survey on Scene Detection Techniques on video

International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2013): 6.14 | Impact Factor (2013): 4.438
National Conference on Knowledge, Innovation in Technology and Engineering (NCKITE), 10-11 April 2015
Kruti Institute of Technology & Engineering (KITE), Raipur, Chhattisgarh, India
Licensed Under Creative Commons Attribution CC BY
A Survey on Different Video Scene Change
Detection Techniques
Dolley Shukla1
, Chandra Shekhar Mithlesh2
, ,Manisha Sharma3
1
Shri shankaracharya College of Engineering and Technology,
Senior Associate Professor Junwani, Bhilai, Dist-Durg, Chhattisgarh 490020, India
2
Shri shankaracharya College of Engineering and Technology,
Junwani, Bhilai, Dist-Durg, Chhattisgarh 490020, India, csmithlesh1987@gmail.com
3
Bhilai Institute of Technology, Professor & Head ( E & Tc )
Durg, Dist-Durg, Chhattisgarh 490020, India, Manihasharma1@rediffmail.com
Abstract: A number of automated scene change detection methods for indexing a video sequence to facilitate browsing and retrieval
have been proposed in recent years. Scene change detection plays an important role in a number of video applications, including video
indexing, semantic features extraction, multimedia information systems, Video on demand (VOD), digital TV, online
processing(networking) neural network mobile applications services and technologies, cryptography, and in watermarking. for copy
protection watermarking is an important technique. Recent advances in technology have made tremendous amount of multimedia
content available. The amount of video is increasing, due to which the systems that improve the access to the video is needed. A current
research topic on video includes video abstraction or summarization, video classification, video annotation and content based video
retrieval. In all these applications one needs to identify shot and key frames in video which will correctly and briefly indicates the
contents of video.
Keywords: Video, Video Sequence, False Detection, Detection Rate, SCD, Recall & Precision, PSNR, Macroblocks etc.
1. Introduction
Video is arguably the most popular means of communication
and entertainment [3]. Video is an important component of
the collaboration data streams. In order to enable filtering of
on-line streams as well as effective retrieval of desired data
from an archive, the video/audio streams need to have
annotations or tags generated that describe the “content” of
the stream[5]. To create indexed video databases, video
sequences are first segmented into scenes, a process that is
referred to as scene change detection(SCD) [1]. With the
rapid growth of video indexing, editing and transcoding
system, video analysis and segmentation become more
important than ever before. Scene change detection plays an
indispensable role in segmenting a long video sequence into
smaller unit for further processing.
A video shot is a sequence of successive frames with similar
visual content in the same physical location. Therefore, abrupt
scene change can be identified when there is a large amount
of visual content change across two successive frames while
gradual transition is much difficult to detect since the
successive frame difference during transition is substantially
reduced[7].
2. Terms of Scene Detection
Image:- Images may be two-dimensional, such as photograph ,
screen display, and as well as a three-dimensional, such as
hologram. They may be captured by optical such as cameras,
mirror, lenses, telescope etc.
Picture Element :- The pixel (a word invented from "picture
element") is the basic unit of programmable color on a computer
display or in a computer image
Video:- A typical video sequence its organized into a
sequence of group of pictures(GOP).A video can be broken
down in scene, shot and frames. each GOP consists of one I
(intra-coded) and a few
P(predicted) and B (bi-directionally interpolated) frames as
shown in Figure 1
DISPLAY ORDER
B
P
P
P
B
B
Figure 1: Group Of Picture (GOP)
I-Picture:-A I-picture is intra-frame coded and does not
depend on any previous or future frames. The bit-rate of a I-
picture depends on the visual content of the picture. If the
picture is of high activity, then the bit-rate will be high.
When a scene change occurs between two consecutive I-
pictures, the content of the two I-pictures will be totally
different. The transition may be from high-activity to low-
activity, low-activity to high-activity, or with similar activity.
Therefore, the bit-rate of the I-picture following the scene
change will be increased, decreased or remained similarly.
This implies that we cannot guarantee to detect all cut points
214

if we only look for the significant change in total bit-rate of
two consecutive I-pictures. However, when two pictures are
of different content, they will have different local statistics
even if their global characteristics are similar. Therefore, if
we measure the bit-rate difference at macroblock level, we
can extract the change of image content between two
consecutive frames of sequence. A large change in bit-rate at
macroblock level between two consecutive I-pictures means
that there is a scene change between them.
P-Picture :-P-picture is a forward motion-predicted frame,
which depends on the previous P- or I-picture. When a scene
change occurs between two consecutive P-pictures, it is
difficult to make the forward prediction since the content of
the current frame will be totally different from the previous
one. Most of the macroblocks have so large prediction error
that they will be encoded using the intra-coded mode. This
will of course lead to a significant increase in the bit-rate of
the current frame. Therefore, we can detect the scene change
by checking whether there is a large increment in the bit-rate
of two consecutive P-frames.
B-Picture :-B-picture is a bidirectional-predicted frame,
which depends on both the previous and future P- or I-picture.
When a scene change occurs, the content of the current frame
will have large difference with the previous anchor frame but
have similar content with the future one. Therefore, most of
the macroblocks depend mostly on the future P- or I-picture
and will be backward predicted. Since the prediction can only
be done mostly in one direction, the coding efficiency of the
motion compensation will be lower and it will lead to the
increase in the bit-rate when compared with the previous B-
picture. However, such a increment will not be as significant
as that occurred in I- or P-picture.
Scene :-A scene is a logical grouping of shots into a semantic
unit.
Shot :-A shot is a sequence of frames captured by a single
camera in a single continuous action. A shot boundary is the
transition between two shots.
Frame:- A digital video consists of frames that are a single
frame consists of pixels.
Precision and Recall :-"Precision" defines how reliable the
detected by the algorithm is while "Recall" defines the overall
performance of the algorithm. Result are reported using F-
measure which is combine precision and recall in a single
measure. These measure are computed by
𝑅𝑒𝑐𝑎𝑙𝑙 =
𝑐𝑜𝑟𝑟𝑒𝑐𝑡
𝑐𝑜𝑟𝑟𝑒𝑐𝑡 + 𝑓𝑎𝑙𝑠𝑒
𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 =
𝑐𝑜𝑟𝑟𝑒𝑐𝑡
𝑐𝑜𝑟𝑟𝑒𝑐𝑡 + 𝑚𝑖𝑠𝑠𝑒𝑑
𝐹 = 2 ×
𝑝𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 × 𝑟𝑒𝑐𝑎𝑙𝑙
𝑝𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛 + 𝑟𝑒𝑐𝑎𝑙𝑙
PSNR :Peak signal-to-noise ratio is the important parameter
of scene change detection algorithm, efficient SCD algorithm
having the value of PSNR greater than 35. All of the SCD
techniques gives approximately same PSNR result or increase
PSNR by at least same.
3. Review on Scene Change Detection
The survey of various research papers that have contributed to
solve the problem of scene change detection in video
sequences, there is a growing demand of video indexing,
scene browsing and retrievals in Signal Process.
The comparison and analysis of scene change detection
algorithms is described in the following subsections
Ralph M. Ford et al.(1997) presented a five metrics for scene
change detection in video sequences. Metric previously
applied to this problem are surveyed, and are quantitatively
compared to the new metrics. The proposed metrics are
superior. These metrics were also applied to the detection of
gradual transitions. y is a good global metric for detecting
abrupt cuts. [1]
Limitations
 The F-test and vi-3 perform the best overall for abrupt cuts,
but require more computation time.
 Global metric did not perform well for gradual transitions,
but the statistic based metrics did.
Taehwan Shin et al.(1998) presented a hierarchical scene
change detection in an mpeg-2 compressed video sequence.
In this paper, we propose an efficient scene change detection
algorithm for direct processing of MPEG-2 video bitstreams.
The proposed algorithm utilizes the hierarchical structure of
the compressed bitstreams and statistical characteristics of the
coded parameters, thus greatly reducing computational
requirement compared to pixel domain processing with full
decompression. Occurrence of scene change is checked first
in a COP level, and if the result is affirmative it is checked
again in lower levels : sub-GOP and each picture.We used
several metrics for different levels : variance of DC images
for I-pictures, number of macroblock types for P-pictures and
motion vector types for B pictures. The proposed algorithm
uses such features of the video which are easily extracted by
minimal decoding of the input bitstream that very efficient
processing is achieved. [2]
Limitations
 In this algorithm, scene change is checked in a hierarchical
fashion from GOP to sub-GOP and picture level, thus
reducing much of the processing requirements.
W.A.C. Fernando et al. (1999) presented a novel algorithm for
wipe scene change detection in video sequences. In the proposed
scheme, each image in the sequence is mapped to a reduced
image. Then statistical features and structural properties of the
images were used to identify wipe transition region. Finally,
Hough transform is used to analyze the wiping pattern and the
direction of wiping .The algorithm is capable of detecting all wipe
regions accurately even when the video sequence contains other
special effects like fading, dissolving, panning the proposed
algorithm can be used in uncompressed video to detect wipe
regions with a very high reliability. [3]
215

Limitations
 Further work is required to extend this algorithm for
compressed video.
W.A.C. Fernando, et al. (2000) presented a unified approach to
scene change detection in uncompressed.There is an urgent need
to extract key information automatically from video for the
purposes of indexing, fast retrieval and scene analysis. To support
this vision, reliable scene change detection algorithms must be
developed. Results on video of various content types are reported
and validated with the proposed scheme in uncompressed and
MPEG-2 compressed video. The accuracy of the detected
transitions is above 95% and 90% for uncompressed and MPEG-2
compressed video respectively. The proposed algorithm can be
used in uncompressed and compressed video to detect scene
changes with a high reliability. The algorithm is capable in of
detecting all scene changes accurately. [4]
Limitations
 Future work is required to extend the algorithm for camera
movements detection within the same framework.
Wensheng Zhou, et al.(2000) proposed an On-line Scene
Change Detection of Multicast Video. Network-based
computing is becoming an increasingly important area of
research, whereby computational elements within a
distributed infrastructure process/enhance the data that
traverses through its path. These computations as online
processing and this paper investigates scene change detection
in the context of MB one based proxies in the network. On-
line processing varies from traditional offline processing
schemes, where for example, the whole video scope is
known as a priori, which allows multiple scans of the stored
video files during video processing. The proposed algorithms
do scene change detection and extract key frames from video
bitstreams sent through the mbone network. Several
algorithms based on histogram differences and evaluate them
with respect to precision, recall, and processing latency a few
effective methods of on-line video scene change detection
over MBone video for the purposes of annotation/filtering.
Main advantages of our algorithms is their ability to support
real-time video processing on the network.Joint algorithms
based on video codec characteristics for acquiring fast and
accurate scene detection were carried out. Both global color
histograms were extracted and block information of DCT
coding was used.. Algorithms were capable of satisfying on-
line annotation needs. The algorithm was tested on different
data bandwidths & gives the best performance in each
case.[5]
Limitations
 Algorithms were mainly targeting real-time video scene
analysis, improving processing speed and reducing latency
are two more issues of concern, which makes it improper to
compare the proposed algorithm with other off-line scene
detection techniques, such as those mentioned in the
Related Work section, where video data are processed off-
line.
 In the future, they plan to study more video on-line
annotations based on video content, such as key frame
classification. Because of the complicated nature of video
processing, they plan on using multiple workstations
working in parallel to realize more sophisticated real-time
annotation
 Plan to study algorithms that are tolerant of packet losses in
the network.
 The on-line processing system is designed to meet the
requirements of real-time video multicasting over the
Internet and to utilize the successful video parsing
techniques available today.
Shu-Ching Chen,et al.(2002 ) presented a change detection
by audio and video clues. Automatic video scene change
detection is a challenging task. Using audio or visual
information alone often cannot provide a satisfactory
solution. However, how to combine audio and visual
information efficiently still remains a difficult issue since
there are various cases in their relationship due to the
versatility of videos. Shu-ching presented an effective scene
change detection method that adopts the joint evaluation of
the audio and visual features. First, video information is used
to find the shot boundaries. Second, the audio features for
each video shot can be extracted. Lastly, an audio-video
combination schema is proposed to detect the video scene
boundaries. Unlike the traditional methods that first analyze
audio and video data separately and then combine them, we
analyze them at different phases The audio feature extraction
is based on the detected video shots, which tends to be more
stable and more reliable in charactering the audio data. The
experimental results demonstrate that our method performs
very well in terms of precision and recall value.[6]
Xiaoquan Yi et al.(2005) presented a Fast Pixel-Based
Video Scene Change Detection. This paper proposes a new
simple and efficient method to detect abrupt scene change
based on only pixel values. Conventional pixel-based
techniques can produce a significant number of false
detections and missed detections when high motion and
brightness variations are present in the video. To increase
scene change detection accuracy yet maintaining a low
computational complexity, a two-phase strategy is developed.
Frames are firstly tested against the mean absolute frame
differences (MAFD) via a relaxed threshold, which rejects
about 90% of the non-scene change frames. The rest 10% of
the frames are then normalized via a histogram equalization
process. A top-down approach is applied to refine the
decision via four features: MAFD and three other features
based on normalized pixel values - signed difference of mean
absolute frame difference (SDMAFD*), absolute difference
of frame variance (ADFV*), and mean absolute frame
differences (MAFD*) method contributes to higher detection
rate and lower missed detection rate while maintaining a low
computational complexity, which is attractive for real-time
video applications. Method is relatively immune from sharp
illumination changes, moving objects, camera motion, and
other similar effects because of the well combined metrics.
Furthermore, our method uses only frame pixel values
without any motion estimation processes so that frugal
computational complexity is maintained, which makes it very
attractive for real-time video applications. [7]
Limitations
 This technique was simple but it performs poorly in cases
of high motions of objects, high motions of cameras, or
severe scene luminance variations (i.e., flicker)
216

Gao, J. Jiang et al.(2006) presented a PCA-based approach
for video scene change detection on compressed video. An
automatic, real-time detection approach to video scene
change detection is presented. Owing to the high correlation
of two consecutive video frames, it is proposed that only the
eigenvector corresponding to the largest eigenvalue is
retained in the principal component analysis (PCA) for video
data. A one-dimensional PCA feature of video data is then
generated from the PCA. It shows superior performance
compared to the histogram feature and the pixel feature.The
detection algorithm based on this PCA feature is then
designed to detect both abrupt and gradual transitions. The
proposed approach is tested on the TREC video test
repository to validate its performance. The superior
performance of the new PCA feature. Then our change
detection algorithm is based on the PCA feature to detect the
scene change. [8]
Limitations
 Tested on the TREC video test repository to validation for
its performance.
J.-R. Ding et al. (2007) presented adaptive group-of-pictures
and scene change detection methods based on existing h.264
advanced video coding information. The H.264 advanced
video coding (H.264/AVC) standard provides several
advanced features such as improved coding efficiency and
error robustness for video storage and transmission. In order
to improve the coding performance of H.264/AVC, coding
control parameters such as group-of-pictures (GOP) sizes
should be adaptively adjusted according to different video
content variations (VCVs), which can be extracted from
temporal deviation between two consecutive frames. The
authors present a simple VCV estimation to design adaptive
GOP detection (AGD) and scene change detection (SCD)
methods by using the obtained motion information, where the
motion vectors and the sum of absolute transformed
differences as VCV features are effectively used to design the
AGD and SCD algorithms, respectively. In order to avoid
unnecessary computation, the above VCV features were
obtained only in the 4 x 4 inter-frame prediction mode.The
proposed AGD with SCD methods can increase the peak
signal-to-noise ratio by 0.62 dB on average over the
H.264/AVC operated with a fixed GOP size. Besides, the
proposed SCD method reached a scene change detection rate
of 98% .On the other hand, if the scene-changed frame
without any detection mechanism is coded by interceding . It
will become inefficient and waste considerable computation
time in motion estimation. and founded that the GOP size for
H.264/AVC should be appropriately decreased and increased
for higher and lower VCV frames, respectively, to improve
the coding performance. The method successfully utilized the
existing video coding information, such as motion vectors and
motion residuals, to become an effective VCV feature. Both
the proposed AGD and SCD methods can effectively help to
adjust a better GOP size to improve the coding performance of
H.264/AVC. the proposed joint AGD and SCD schemes can
increase PSNR by at least 0.62 dB compared with the
H.264/AVC operated in the fixed GOP size on average. [9]
Limitations
 The proposed and some existing VCV characteristics can
be further extended for adaptive search window, adaptive
search methods (of three step search and
ZHANG Ji et al.(2007) presented an effective error
concealment framework for h.264 decoder based on video
scene change detection. In this paper, we propose an effective
error concealment framework for H.264 decoder based on the
scene change detection. The proposed framework quickly
and accurately detects whether scene change occurs in the
decoding frame, based on the detection result, both corrupted
intra frames and damaged inter frames can be reconstructed
by spatial or improved temporal EC (Error Concealment)
algorithm. The experiment shows that, compared with the
traditional error concealment method in the H.264/A VC non
normative decoder, the proposed framework has better
robustness and can efficiently improve the visual quality and
PSNR of the decoded video. [10]
Limitations
 The final experimental results demonstrate that our
framework can apparently improve the visual quality of
reconstructed video sequence and PSNR.
 In the future, they will emphasize more on exploring the
scene change algorithm for B frames and improve spatial
or temporal EC algorithms in the new frame work.
Jens Brandt et al.(2008) presented a fast frame-based scene
change detection in the compressed domain for mpeg-4
video. Detection of scene changes is an elementary step in
automatic video processing like indexing, segmentation or
transcoding Video indexing and segmentation allow fast
browsing without decoding the complete video. In the case of
transcending, the information about scene changes such as
cuts and fades as well as about special movements like
rotations or zooms in video frames is helpful to determine
suitable transcoding parameters. In the compressed domain
only information about DCT values as well as motion
information can be used to determine such scene changes and
movements. Therefore different measures were defined
which use the encoded DCT values and motion vectors of
each compressed frame. Based on these measures as well as
on motion vector histograms, a fast approach to detect
different kinds of scene changes and special movements in
MPEG-4 videos in the compressed domain was proposed. In
this paper we have presented a fast frame-based algorithm for
compressed domain scene change detection in MPEG-4
video streams. The easy computation of the used metrics and
histograms makes the whole scene change detection
algorithm very fast which is very important for real time
video processing. Therefore the scene change detection can
only be a tool to indicate a stronger or weaker correlation
towards one or the other scene movement. However, the
evaluation results show that the algorithm detects a high
number of scene changes and movements. Based on these
promising results Jens Brandt developed a method to use the
information about detected scene changes and special
movements within a video stream for determining suitable
transcoding parameters. [11]
217

Limitations
 The highly statistical nature of the scene change detection
results in a statement that by design cannot be absolutely
correct.
Hakan Haberdar et al.(2010) presented a disparity map
refinement for video based scene change detection using a
mobile stereo camera platform. This paper presents a novel
disparity map refinement method and vision based
surveillance framework for the task of detecting objects of
interest in dynamic outdoor environments from two stereo
video sequences taken at different times and from different
viewing angles by a mobile camera platform. Preliminary
disparity images were refined based on estimated disparities
in neighboring frames. Segmentation is performed to
estimate ground planes, which in turn are used for
establishing spatial registration between the two video
sequences. Finally, the regions of change are detected using
the combination of texture and intensity gradient features.
We present experiments on detection of objects of different
sizes and textures in real videos. The framework is evaluated
on 4 real video sequences and the experimental results show
that the proposed framework is able to detect changes/objects
of different sizes and textures, and is able to ascertain if there
are no changes in the scene, thereby minimizing false
positives. [12]
Limitations
 This method need for future work and it will focus on
relaxing our assumption on known temporal alignment and
improving the spatial registration module.
Ufuk Sakarya et al.(2010) presented a video scene detection
using graph-based representations. One way of realizing this
step is to obtain shot or scene information. One or more
consecutive semantically correlated shots sharing the same
content construct video scenes. On the other hand, video
scenes are different from the shots in the sense of their
boundary definitions; video scenes have semantic boundaries
and shots are defined with physical boundaries. In this paper,
we concentrate on developing a fast, as well as well-
performed video scene detection method. Our graph partition
based video scene boundary detection approach, in which
multiple features extracted from the video, determines the
video scene boundaries through an unsupervised clustering
procedure. For each video shot to shot comparison feature, a
one-dimensional signal is constructed by graph partitions
obtained from the similarity matrix in a temporal interval.
After each one-dimensional signal is filtered, an unsupervised
clustering is conducted for finding video scene boundaries.We
adopt two different graph-based approaches in a single
framework in order to find video scene boundaries. The
proposed graph-based video scene boundary detection method
is evaluated and compared with the graph-based video scene
detection method presented in literature. [13]
Limitations
 Some issues are considered into account for future research
directions. Different video features can be used in order to
improve the results and; moreover, content varying based
changes in the multiple features may increase the scene
detection performance.
 Especially, feature weighting in clustering process can
also be applied. In addition, an adaptation of the proposed
method for other types of videos, like news,
documentaries, etc., is another research direction.
Ankita P. Chauhan1 et al (2013) presented Hybrid
Approach for Video Compression Based on Scene Change
Detection. In order to fulfill the requirement of limited
channel bandwidth and of growing video demand like
streaming media delivery on internet, and digital library,
video compression is necessary. In video compression,
temporal redundancy between adjacent frames is removed
with block based motion estimation algorithms. Video
represents a sequence of frames captured from camera. Scene
is a series of consecutive frames captured from narrative
point of view. In this paper we present an effective scene
change detection method for an uncompressed video. We
have divided frames in to blocks and applied a canny edge
detector in consecutive frames. Count no of pixels (ones) in
each block and compare it with consecutive frames. If scene
change happens then number of pixels per block will change,
based on that change we can detect scene change in
consecutive frame. presented a hybrid approach, in which we
have used scene change detection along with block based
motion estimation algorithms (BME) to compress video.In
this paper we have presented a simple approach with high
accuracy. To reduce the number of computations, hybrid
approach along with scene change detection have been
implemented.[14]
Limitations
 Simulation result shown that Proposed Hybrid approach
compared with NTSS, 4SS and DS algorithm, it reduces
the cost of computational complexity and also gives
approximately same PSNR result.
Neetirajsinh J. Chhasatia et al. (2013) presented a
Performance Evaluation of Localized Person Based Scene
Detection and Retrieval in Video. A work regarding person
localization using novel approach had been carried out.
Person localization is the first step in all the surveillance
application in multimedia and human-computer interface
applications. The objective of was to address a system for a
person localization system which described important tasks
of video content analysis, namely, video segmentation,
localizing the person and their identification. A novel
framework was proposed for combined video segmentation
and retrieval of scene with object/person localization. In
which the dominant regions as a large area of the frame have
been tracked throughout a shot and then stable features were
extracted. Dominant region based effective parameters
(DREP) model which includes low level features of the
proposed model. The low level features were referenced as
the color of dominant region, edges, area, centroid and
position of a particular object or person to be localized in the
video. Based on the extracted information dimensional
processing was used to extract images of faces or the
required objects. The simulated results based on the
presented algorithm shows good performance & highly
robust to camera, objects motions and could withstand
illumination changes even in different poses in the sequences
of the images.[15]
218

Limitations
 Improvement in its basic inputs would result in better
performance of all these techniques.
4. Conclusion
The above survey of various researchers of different
algorithms of Scene Change Detection shows that the two
Main problem associated with xisting algorithms i.e.
threshold-dependent algorithms and color Histogram
Method.
The threshold-dependent algorithm gives false detection
with scenes involving fast camera or object motion. Color
Histogram Method having some advantages like Efficient
representation, Easy computation, Global color distribution,
Insensitive to Rotation, Zooming Changing resolution and
Partial occlusions. Bu some disadvantages also like
Ignorance to spatial relationship, Sensitive in illumination
changes and Chooses illumination-insensitive color
channels. The method avoided the false alarms by using the
validation mechanism. It also proves that the statistical
model-based approach is reliable for gradual scene-change
detection. A very high detection rate is achieved while the
false alarm rate is comparatively low. For better output
detection, a scene change detection algorithm should be
proposed with the properties low false alarm,high
robustness in gradual transition, camera fabrication, special
events and so on . For a considerably good quality of scene
change detection (SCD) algorithm PSNR ranges should be
>40, Recall & precision rate > 95% and change detection rate
>98%.
References
[1] Ralph M. Ford, Craig Robson, Daniel Temple, and
Michael Gerlach “Metrics for Scene Change Detection
in Digital Video Sequences”in IEEE International
Conference on Multimedia Computing and Systems,
Year-1997, IEEE, pp.610 - 611.
[2] Taehwan Shin, Jae-Gon Kim, Hankyu Lee and Jinwoong
Kim “Hierarchical Scene Change Detection In An
Mpeg-2 Compressed Video Sequence” in Circuits and
Systems, Volume-4 , Year-1998, IEEE, pp.253 - 256.
[3] W.A.C. Fernando, C.N. Canagarajah, D. R. Bull “Wipe
Scene Change Detection in Video Sequences” in Image
Processing, Volume-3, Year-1999, IEEE, pp.294-298.
[4] W. A. C. Fernando, C. N. Canagarajah and D. R. Bull
“A Unified Approach To Scene Change Detection In
Uncompressed And Compressed Video ”in Volume-46,
issue-3, Year-2000, IEEE, pp.769-779.
[5] Wensheng Zhou1 and Asha Vellaikal “On-line Scene
Change Detection of Multicast Video” in Journal of
Visual Communication and Image Representation,
Volume-3527, Year-2001, pp.1-15
[6] Shu-Ching Chen, Mei-Ling Shyu, Wenhui Liao,
Chengcui Zhang “Scene Change Detection By Audio
And Video Clues” IEEE. Volume-2,Year-2002 , pp.365
- 368.
[7] Xiaoquan Yi and Nam Ling “Fast Pixel-Based Video
Scene Change Detection" in volume-4 Year-2005, IEEE
, pp. 3443 - 3446.
[8] L. Gao, J. Jiang, J. Liang, S. Wang, S. Yang and Y. Qin
“PCA-based approach for video scene change detection
on compressed video” in Electronics Letters, Volume-
42, issue-24. Year-2006, pp.1389 - 1390 .
[9] J.-R. Ding J.-F. Yang “Adaptive group-of-pictures and
scene change detection methods based on existing H.264
advanced video coding information” in Image
Processing, IET , Volume-2, Issue-2 Year-2008, pp. 85
- 94.
[10] ZHANG Ji, LIU Feng, SHAO Hui, WANG Gang “An
Effective Error Concealment Framework For H.264
Decoder Based on Video Scene Change Detection”
Fourth International Conference on Image and Graphics,
Year-2007, IEEE, pp. 285 - 290.
[11] Jens Brandt Jens Trotzky Lars Wolf “Fast Frame-Based
Scene Change Detection in the Compressed Domain for
MPEG-4 Video ” The Second International Conference
on Next Generation Mobile Applications, Services, and
Technologies, Year-2008, IEEE, pp.514 – 520
[12] Hakan Haberdar and Shishir K. Shah “Disparity Map
Refinement for Video Based Scene Change Detection
Using a Mobile Stereo Camera Platform” International
Conference on Pattern Recognition, Year 2010 IEEE pp.
3890 - 3893.
[13] Ufuk Sakarya and Ziya Telatar “Video scene detection
using graph-based representations” Signal Processing:
Image Communication 25 (2010) Elsevier, Year- 2010 ,
pp. 774–783
[14] Ankita P. Chauhan Rohit R.Parmar, Shankar K.Parmar
Andshahida G.Chauhan “Hybrid Approach for Video
Compression Based on Scene Change Detection” in
Signal Processing, Computing and Control (ISPCC),
Year-2013, IEEE, pp.1-5
[15] Neetirajsinh J. Chhasatia, Chintan U. Trivedi and Komal
A. Shah “Performance Evaluation of Localized Person
Based Scene Detection and Retrieval in Video ” Image
Information Processing (ICIIP), 2013 IEEE Second
International Conference on Year- 2013 , pp. 78 – 8
Author Profile
Chandra Shekhar Mithlesh received
the B.E. degrees from government
Engineering College jagdalpur in
Electronics & Telecommunication with
first Class and currently perusing his
master degree in Communication
Engineering from Shri shankaracharya
College of Engineering and Technology,, Bhilai,, Chhattisgarh.
219

survey on Scene Detection Techniques on video

More Related Content

What's hot (20)

Similar to survey on Scene Detection Techniques on video (20)

Recently uploaded (20)

survey on Scene Detection Techniques on video