Some Research Topics and Projects (Past and Present)
Focus on
Image Processing, Computer Vision and Pattern Recognition |
Image Denoising and
Enhancement (software available):
- Based
on wavelet decomposition
- Adaptive shrinkage or
enlargement of wavelet coefficients
- Some applications in
mammographic and dermatological images
Original
Image
Noisy
Image
Denoised Image
Related publications (click here for the full publications list)
- JUNG, Cláudio
Rosito; SCHARCANSKI, Jacob. Adaptive Image Denoising
in Scale-Space Using the Wavelet Transform. In: SIBGRAPI, 2001, Florianópolis
- SC. Proceedings of SIBGRAPI 2001 (XIVI Brazilian Symposium on
Computer Graphics and Image Processing). IEEE Computer Society, 2001.
p. 172-178.
- JUNG, Cláudio
Rosito; SCHARCANSKI, Jacob. Wavelet-based
method for image filtering using scale-space continuity. Journal of
Electronic Imaging, v. 10, n. 2, p. 511-520, 2001.
- SCHARCANSKI, Jacob; JUNG, Cláudio Rosito;
CLARKE, Robin T. Adaptive Image Denoising Using Scale and Space
Consistency. IEEE Transactions on Image Processing, v. 11, n. 9, p.
1092-1101, 2002.
- JUNG, Cláudio
Rosito; SCHARCANSKI, Jacob. Adaptive image denoising
and edge enhancement in scale-space using the wavelet transform. Pattern
Recognition Letters, v. 24, n. 7, p. 965-971, 2003.
- JUNG, Cláudio
Rosito; SCHARCANSKI, Jacob. Wavelet
transform approach to adaptive image denoising and
enhancement. Journal of Electronic
- SCHARCANSKI, Jacob, JUNG, Cláudio Rosito Denoising
and Enhancing Digital Mammographic
Images for Visual Screening. Computerized Medical Imaging And Graphics.
, v.30, p.243-254, 2006
- JUNG, Cláudio
Rosito, SCHARCANSKI, Jacob. Sharpening
Dermatological Color Images in the Wavelet Domain. IEEE Journal of
Selected Topics in Signal Processing. , v.3, p.4 - 13, 2009.
Multiscale Image
Segmentation (software available):
- Multiscale image decomposition
based on the wavelet transform
- Detection of relevant edges by
analyzing detail coefficients
- Segmentation
based on controlled watersheds
Original
Image
Enhanced
Image
Segmented Image
Original
Image
Noisy
Image
Segmented Image
Related publications (click here for the full publications list)
- JUNG, Cláudio
Rosito; SCHARCANSKI, Jacob. Robust Watershed
Segmentation Using the Wavelet Transform. In: SIBGRAPI, 2002,
- JUNG, Cláudio Rosito. Multiscale Image
Segmentation Using Wavelets and Watersheds. In: SIBGRAPI, 2003, São
Carlos, SP. Proceedings of SIBGRAPI 2003 (XVI Brazilian Symposium on
Computer Graphics and Image Processing). Los Alamitos, USA: IEEE Computer Society, 2003. v. 1, p. 278-284.
- JUNG, Cláudio
Rosito; SCHARCANSKI, Jacob. Robust Watershed
Segmentation Using Wavelets. Image and Vision Computing, v. 13, n. 7,
p. 661-669, 2005.
- JUNG, Cláudio Rosito. Combining Wavelets and
Watersheds for Robust Multiscale
Image Segmentation. Image and Vision Computing. , v.25, p.24 – 33,
2007.
Vision-based Driver
Assistance Systems:
- Linear-parabolic lane model is
adopted in image coordinates
- Linear portion models the near
field, allowing and estimate of the lane orientation
- Parabolic portion models the
far field, allowing the detection of incoming curves
Lane departure detection: red
lane borders indicate lane departure. DivX movie available (~1MB) – See movie
- Self
vehicular onboard camera calibration
- Explores the expected lane
geometry to estimate some extrinsic camera parameters
- Intrinsic parameters must be
known (easy to calibrate offline)
- Possible applications involving
Augmented Reality (AR)
- For some videos, check http://www.inf.ufrgs.br/~crjung/vehicle_ar/
- Detection and classification of horizontal lane marking
- Explores spatial and temporal characteristics of the contrast between lane markings and the pavement
- Dataset avaliable at http://www.inf.ufrgs.br/~crjung/vehicle_ar/roadLaneMarkings
Related publications (click here for the full publications list)
- de
PAULA, Maurício B. ; JUNG, Cláudio Rosito . Automatic Detection and
Classification of Road Lane Markings Using Onboard Vehicular Cameras.
IEEE Transactions on Intelligent Transportation Systems (Print), p.
1-10, 2015.
- de PAULA,
Maurício B. ; JUNG, Cláudio Rosito ; da SILVEIRA JR., Luiz G. . Automatic on-the-fly extrinsic camera calibration
of onboard vehicular cameras. Expert Systems with Applications, v. 41, p.
1997-2007, 2014.
- de PAULA, Mauricio B.; JUNG, Claudio R. Real-time
detection and classification of road lane markings. In Conference on
Graphics, Patterns, and Images, 2013, p. 83-90;
- JUNG, Cláudio Rosito; KELBER, Christian
Roberto A Robust Linear-Parabolic Model for Lane Following In: SiBGRAPI,
2004,
- JUNG, Cláudio Rosito; KELBER, Christian
Roberto. Lane Following and Lane Departure using a Linear-Parabolic Model.
Image and Vision Computing, v. 23, n. 13, p. 1192-1202, 2005.
- JUNG, Cláudio Rosito; KELBER, Christian
Roberto. A Lane Departure Warning System Using Lateral Offset with Uncalibrated
Camera. In: IEEE INTELLIGENT TRANSPORTATION SYSTEMS CONFERENCE, 2005, Viena,
Áustria.
Proceedings of the IEEE Intelligent Transportation Systems Conference. 2005.
p. 348-353.
- JUNG, Cláudio
Rosito; KELBER, Christian Roberto. An Improved Linear-Parabolic Model for
Lane Following and Curve Detection. In: SIBGRAPI, 2005,
Background
Subtraction with Shadow Removal (software
available):
- Statiscal model for static
background pixels (coping with moving objects during training)
- Simple/fast foreground test
(concurrent implementations)
- Shadow/highlight detection
based on properties of pixel ratios
Related publications (click here for the full publications list)
- JACQUES JR,
Julio Cezar Silveira; JUNG, Cláudio Rosito; MUSSE, Soraia Raupp.
Background Subtraction and Shadow Detection in
Grayscale Video Sequences. In: SIBGRAPI, 2005,
- JACQUES JR, Julio Cezar Silveira,
JUNG, Cláudio
Rosito, MUSSE, Soraia
Raupp
A Background Subtraction Model Adapted to Illumination Changes In: IEEE
International Conference on Image Processing, Atlanta, USA. Proceedings
of the IEEE International Conference on Image Processing. IEEE Press,
2006. p.1817 – 1820
- JUNG, Cláudio Rosito. Efficient Background
Subtraction and Shadow Removal for Monochromatic Video Sequences. IEEE
Transactions on Multimedia, v. 11, p. 571-577, 2009.
Wavelet-based
Texture Synthesis and Image Inpainting:
- Texture
Synthesis
- Explore wavelets for
block-based texture synthesis
- Inpainting
- Decimated wavelet tranform for
smoothing and edge detection
- Block-based matching based on
the coherence of wavelet coefficients in the overlap region
- Blocks
with different sizes
|
|
|
|
Texture Synthesis |
Image Inpainting (click here for video) |
Related publication (click here for the full
publications list)
- TONIETTO,
Leandro ; WALTER, Marcelo ; JUNG, Cláudio Rosito . Patch-Based Texture Synthesis using
Wavelets. In: Brazilian Symposium on Computer Graphics and Image
Processing (SIBGRAPI), 2005, Natal, RN. Proceedings of SIBGRAPI 2005
(XVIII Brazilian Symposium on Computer Graphics and Image Processing). Los
Alamitos, USA: IEEE Computer Press, 2005. v. I. p. 383-389
- TONIETTO,
Leandro ; WALTER, Macelo ; JUNG, Cláudio Rosito . A Randomized Approach for Patch-Based
Texture Synthesis using Wavelets. Computer Graphics Forum,
v. 25, p. 675-684, 2006.
- IGNÁCIO,
Ubiratã Azevedo; JUNG, Cláudio Rosito. Block-Based Image Inpainting in the Wavelet Domain. The
Visual Computer, v.23, p.733 - 741, 2007
Object Tracking,
focused on pedestrian and face (software available):
- Global motion based on the
motion of points/patches
- Rectangular pacthes used for
pedestrian tracking
- When calibrated cameras are
available, scale adaptation is done implicitly
- Lucas-Kanade based tracking
for face tracking
- Translational
and scale information
- Combination of individually
tracked patches/points using Weighted Vector Median Filters
- Provide an average of individual
motion vectors implicitly accounting for outliers
- Use of camera parameters for
imposing geometrically coherent paths in pedestrian tracking
|
|
Click here to see a vídeo (~1MB) |
We also provide some original videos with
ground truth data for tracking algorithms related to one of our trackers. See
the CPD Tracker
page for more details.
For pedestrian tracking using calibrated
cameras, please go to Gustavo Fuhr's research page
Related publications (click here for the full publications list)
- JUNG, Cláudio Rosito ; SAID, Amir
. Object Tracking Using Multiple Fragments. Proceedings of the IEEE
16th International Conference on Image Processing, 2009. p. 889-892.
- BINS, Jose C. ;
JUNG, Cláudio Rosito ; DIHL, Leandro L. ; SAID, Amir . Feature-based face tracking for videoconferencing
applications. In: IEEE International Symposium on Multimedia, 2009,
- DIHL, Leandro L.
; JUNG, Cláudio Rosito ; BINS, Jose C. . Robust Adaptive Patch-based Object Tracking using Weighted Vector
Median Filters. In: Conference on Graphics, Patterns and Images, 2011,
Maceio. Conference on Graphics, Patterns and Images, 2011. p. 149-156.
- FUHR, Gustavo ;
JUNG, CLÁUDIO R. . Robust
Patch-Based Pedestrian Tracking using Monocular Calibrated Cameras. In:
Conference on Graphics, Patterns and Images, 2012, Ouro Preto, MG.
Proceedings of XXV SIBGRAPI - Conference on Graphiocs, Patterns and
Images, 2012. p. 166-173.
- BINS, Jose C. ; DIHL, Leandro L. ; JUNG, Cláudio Rosito. Target Tracking using Multiple Patches and Weighted Vector Median Filters. Journal of Mathematical Imaging and Vision, v. 45, p. 293-307, 2013.
- FUHR, Gustavo ; JUNG, Cláudio Rosito . Combining patch matching and detection for robust pedestrian tracking in monocular calibrated cameras. Pattern Recognition Letters, v. 39, p. 11-20, 2014.
- FUHR, Gustavo ; JUNG, CLAUDIO R. . Camera Self-calibration Based on Non-Linear Optimization and Applications in Surveillance Systems. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, v. 27, p. 1132-1142, 2017.
Detection of
Events/Unusual Motion:
- Given the tracks of the
pedestrians, different approaches to detect events/unusual motion
- Grouping
Behavior
- Voronoi Diagrams (VDs) are
built at each frame, using the positions of the pedestrians as sites
- Temporal evolution of the VDs
as well as psico-social distances used to detect grouping behavior
- Global
trajectory analysis
- Trajectories clustered
according to main flows
- Local information (position
and velocity) used to find the coherence of a test trajectory with each
cluster
- Clusters and local
information can also be used for comparing two crowds
Clustering
in the feature
space
Clusters
of trajectories (outliers in black)
Related publications (click here for the full
publications list)
- JUNG, Cláudio Rosito,
JACQUES JR, Julio Cezar Silveira,
SOLDERA, John, MUSSE, Soraia Raupp Detection
of Unusual Motion Using Computer Vision In: SiBGRAPI,
2006, Manaus. Proceedings
of XIX Brazilian Symposium on Computer Graphics and Image Processing. Los Alamitos, CA: IEEE Computer Society,
2006. v.I. p.349
– 356
- JACQUES JR,
Julio Cezar
Silveira
; BRAUN, Adriana ; SOLDERA, John ; MUSSE, Soraia Raupp
; JUNG, Cláudio
Rosito . Understanding People Motion in Video Sequences
Using Voronoi Diagrams. Pattern Analysis and Applications, 2007,
v.10, p.321 - 332, 2007.
- JUNG, Cláudio
Rosito, HENNEMANN, Luciano, MUSSE, Soraia Raupp. Event Detection
Using
Trajectory
Clustering
and
4D
Histograms.
IEEE Transactions on Circuits An Systems for Video Technology. ,
v.18, p.1565 - 1575, 2008.
- MUSSE, Soraia R. ; CASSOL, Vinicius ; JUNG, Cláudio Rosito. Towards
a Quantitative Approach for Comparing Crowds. Computer Animation and Virtual Worlds
(print), v. 23, p. 49-57, 2012.
- ALMEIDA, Igor; JUNG, Claudio
R.; Change detection in human crowds. In Conference on Graphics, Patterns, and Images, 2013,
p. 64-39.
Audio and/or Video-based Approaches for Voice Activity
Detection (VAD) and Sound Source Localization (SSL)
- Video-based
approaches
- Detect/track faces in video
sequences
- Extract the temporal evolution
of the lips regions for VAD
- Audio-based
approaches
- Use of an array of
omnidirectional microphones
- Explore the spatio-temporal
consistency of the detected sound source based on the SRP-PHAT algorithm
- GPU
Implementation for improved performance
- Audiovisual
approaches
- Fusion of audio and visual
cues to improve VAD and SSL
- Focus on methods that
detect simultaneous speech
- Inclusion of depth information
(RGB-D cameras)
|
|
|
|
Visual VAD (Based on lip
motion) |
Audiovisual VAD (click here for video) |
Related publications (click here for the full
publications list)
- SCOTT, Dario
; JUNG, Cláudio Rosito ; BINS, Jose C. ; SAID, Amir ; KALKER, Antonius .
Video based VAD using adaptive color information. In: IEEE International
Symposium on Multimedia, 2009, San Diego. Proceedings of XI IEEE ISM
International Symposium on Multimedia, 2009. p. 80-87.
- da SILVEIRA JR., Luiz G. ; MINOTTO,
Vicente P. ; JUNG, Cláudio Rosito ; LEE, Bowon . A GPU
Implementation of the SRP-PHAT Sound Source Localization Algorithm. In:
12th International Workshop on Acoustic Echo and Noise control, 2010, Tel
Aviv, Israel. Proceedings of the 12th International Workshop on Acoustic
Echo and Noise control, 2010. p. 1-4.
- LOPES, Carlos B. O. ; GONCALVES,
Andre L. ; SCHARCANSKI, Jacob ; JUNG, Cláudio Rosito . Color-based
lips extraction applied to voice activity detection. In: IEEE
International Conference on Image Processing, 2011, Bruxelas. Proceedings of IEEE International
Conference on Image Processing, 2011. p. 1081-1084.
- BLAUTH, D. ;
MINOTTO, Vicente P. ; JUNG, Cláudio Rosito ; LEE, Bowon ; KALKER, Antonius
. Voice Activity Detection and Speaker Localization using Audiovisual
Cues. Pattern Recognition Letters, v. 33, p. 373-380, 2012.
- MINOTTO, Vicente P. ; LOPES, Carlos
B. O. ; SCHARCANSKI, Jacob ; JUNG, Cláudio Rosito ; LEE, Bowon . Audiovisual
Voice Activity Detection Based on Microphone Arrays and Color Information.
IEEE Journal of Selected Topics in Signal Processing, v. 7, p. 147-156,
2013.
- MINOTTO, Vicente P. ; JUNG, Cláudio
Rosito ; da SILVEIRA JR., Luiz G. ; LEE, Bowon . GPU-based
Approaches for Real-time Sound Source Localization using the SRP-PHAT
Algorithm. The International Journal of High Performance Computing
Applications, v. 27, p. 291-306, 2013.
Stereo/Multiview Stereo Matching and View
Interpolation
- Given two or more rectified
cameras
- Build a 2D triangular
tesselation of the domain
- Perform region-based matching
using the triangular regions
- Refine corners of the mesh to
smooth the disparity map
- Build a 3D mesh by collapsing
coners or inserting orthogonal triangles
- Generate synthetic views using
one or more 3D triangular meshes with texture
|
|
|
|
|
|
Reference image |
Domain tesselation |
Initial disparity map |
Refined disparity map |
Related publications (click here for the full
publications list)
- FICKEL, G. ; JUNG, Cláudio Rosito ;
SAMADANI, R. ; MALZBENDER, T. . Stereo Matching based on Image
Triangulation for View Synthesis. In: IEEE International Conference on
Image Processing, 2012, Orlando. Proceedings of IEEE International Conference on Image
Processing, 2012. p. 2733-2736.
- FICKEL, Guilherme P.,
JUNG, Claudio R., MALZBENDER, Tom, SAMADANI, Ramin., CULBERTSON, Bruce.
Stereo Matching and View Interpolation Based on Image Domain
Triangulation. IEEE Transactions on Image Processing, v.22,
p.3353 - 3365, 2013.
- FUHR,
Gustavo; FICKEL, Guilherme P.; DAL'AQUA, Lorenzo P.; JUNG, Claudio R.;
Malzbender, Tom; Samadani, Ramin. An Evaluation of Stereo Matching Methods for View Interpolation. In
IEEE International Conference in Image Processing, 2013, p. 403-407.
Human Body Segmentation in Still Images (cooperation
with PUC-RS)
- Given a still image (e.g.
photograph), the goal is to extract the silhuette wirth semantic
information
- Our approaches are based on
finding minimum/maximum cost paths in a graph
- The
skeletion is manually informed
- A graph around the skeleton is
build, and an energy function based on edges, anthropometry and coherence
with the corresponding "bone" is used to link the graph edges
- As output, each body part
(arms, legs, etc.) is identified and labeled
|
|
|
|
Silhouette grphs |
Example of body
segmentation (red points separate different body parts) |
Related publications (click here for the full
publications list)
- JACQUES JR, Julio C. S. ; DIHL, Leandro L. ; JUNG, Cláudio
Rosito ; THIELO, Marcelo R. ; KESHET, Renato ; MUSSE, Soraia R. . Human
Upper Body Identification from Images. In: IEEE International Conference
on Image Processing, 2010, Hong Kong. Proceedings of the IEEE
International Conference on Image Processing, 2010. p. 1717-1720.
- JACQUES JR, Julio C. S. ; JUNG, Cláudio Rosito ; MUSSE,
Soraia R. . Skeleton-Based Human Segmentation in Still
Images. In: IEEE International Conference on Image Processing, 2012,
Orlando. Proceedings of
IEEE International Conference on Image Processing, 2012. p. 141-144.
- JACQUES JR., Julio C. S.;
DIHL, Leandro L.; MUSSE, Soraia R.; Self-Occlurion and 3D Pose Estimation
in Still Images. In IEEE
International Conference in Image Processing, 2013, p. 2539-2543.