Special Sessions
Depth camera image processing and applications
Session Organizers: Seungkyu Lee
Session Description: Depth cameras have become increasingly popular for the last couple of years. While depth cameras are very attractive because they provide 3D information instantaneously, they are based on active lighting, e.g., time of flight or structured light. A direct consequence of these active sensing technologies is that 3D cameras are limited in range, have low resolution and unique noises in the presence of scattering or non-Lambertian surfaces. This session covers depth image processing methods such as noise modeling, upsampling and removing range ambiguity as well as applications such as tracking, segmentation, view synthesis and lighting estimation.
Papers
- 3470: Seungkyu Lee, "DEPTH CAMERA IMAGE PROCESSING AND APPLICATIONS"
- 2222: Amira Belhedi, Steve Bourgeois, Vincent Gay-bellile, et al., "NON-PARAMETRIC DEPTH CALIBRATION OF A TOF CAMERA"
- 2356: Daeyoung Kim and Kuk-jin Yoon, "HIGH QUALITY DEPTH MAP UP-SAMPLING WITH CONSIDERATION FOR EDGE NOISE OF RANGE SENSORS"
- 1437: Ouk Choi and Seungkyu Lee, "Wide range stereo time-of-flight camera"
- 3070: Fei Yang, Junzhou Huang et al., "ROBUST FACE TRACKING WITH A CONSUMER DEPTH CAMERA"
- 3304: Huynjung Shim, "ESTIMATING ALL FREQUENCY LIGHTING USING A COLOR/DEPTH IMAGE"
- 1612: Haowei Liu, Matthai Philipose and Ming-Ting Sun, "AUTOMATIC OBJECT SEGMENTATION WITH RGB-D CAMERAS"
- 3102: Anjin Park and Jinwook Kim, "GPU ACCELERATED VIEW SYNTHESIS FROM MULTIPLE RGB-D IMAGES"
People re-identification and tracking from multiple cameras
Session Organizers: Tiziana D'Orazio and Grazia Cicirelli
Session Description: Over the last five years people re-identification and tracking have been gaining increasing attention in the computer vision community. This task refers to the problem of recognizing an individual captured in different times and locations across spatially non-overlapping cameras. It is very crucial for modern security and surveillance systems, in which wide areas have to be monitored. Re-identification and tracking algorithms have to be robust in challenging situations caused by differences in camera viewpoints and orientations, varying lighting conditions, pose variability of persons, and rapid change in clothes appearance. Recent methods mainly consider appearance based approaches to compute distinctive and invariant signatures for the person being tracked. People signatures can be generated by the analysis of a single image, or by the fusion of information coming from multiple images provided by tracking. Many local/global features and their weighted combinations are used to characterize both colour and texture of each individual, and different classification methodologies are tested to match corresponding signatures.
The special session will be opened by an overview talk and will propose 6 scientific contributions on recent research approaches, discussing open new perspectives on this emerging research field.
Papers
- 3471: Tiziana D'Orazio and Grazia Cicirelli, "PEOPLE RE-IDENTIFICATION AND TRACKING FROM MULTIPLE CAMERAS: A REVIEW"
- 3479: Slawomir Bak, Duc Phu Chau, et al., "MULTI-TARGET TRACKING BY DISCRIMINATIVE ANALYSIS ON RIEMANNIAN MANIFOLD"
- 3477: Luis Teixeira, Pedro Carvalho, et al., "AUTOMATIC DESCRIPTION OF OBJECT APPEARANCES IN A WIDE-AREA SURVEILLANCE SCENARIO"
- 3482: William Schwartz, "SCALABLE PEOPLE RE-IDENTIFICATION BASED ON A ONE-AGAINST-SOME CLASSIFICATION SCHEME"
- 3490: Martin Hirzer, Csaba Beleznai, et al., "DENSE APPEARANCE MODELING AND EFFICIENT LEARNING OF CAMERA TRANSITIONS FOR PERSON RE-IDENTIFICATION"
- 2202: Wei Li, Yang Wu, et al., "COMMON-NEAR-NEIGHBOR ANALYSIS FOR PERSON RE-IDENTIFICATION"
- 2998: Daniel Wesierski, Patrick Horain and Zdzislaw Kowalczuk, "EBE: ELASTIC BLOB ENSEMBLE FOR COARSE HUMAN TRACKING"
Recent advances in cryptography and image processing
Session Organizers: Inald Lagendijk, William Purech and Zekeriya Erkin
Session Description: With the rapid growth of processing power and network bandwidth, many multimedia applications have emerged in the recent past. As digital data can easily be copied and modified, the concern about its protection and authentication have surfaced. Image and video security becomes increasingly important for many applications, e.g., confidential transmission, video surveillance, military and medical applications. For example, the necessity of fast and secure diagnosis is vital in the medical world. Digital rights management (DRM) has emerged as an important research field to protect the copyrighted multimedia data. DRM systems enforce the rights of the multimedia property owners while ensuring the efficient rightful usage of such property. Encryption is used to restrict the access of multimedia content to only authorized users. So far, no encryption scheme has been standardized for image and video encryption and lot of research is being conducted in this domain. Multimedia data requires either full encryption or selective encryption (SE) depending on the application requirements. For example military and law enforcement applications require full encryption while commercial multimedia content required a lower level protection. Such approaches reduce the computational requirements in networks with diverse client device capabilities. SE can be used to process and transmit real-time video data. In case of SE, we have a trade-off between the security level and the necessary computational resources. In addition to SE, security and privacy considerations in image and video processing applications have become increasingly important in recent years. A number of cryptographic solutions have been proposed for controlling access to private image and video content, for protecting the privacy in biometric identification systems (face detection/recognition, fingerprint and iris matching), for privacy preserving data mining and for protecting the private parameters in (commercially) valuable algorithms. The proposed solutions apply a wide range of cryptographic tools, ranging from homomorphic encryptions to secure multi-party computation techniques and differential privacy, on to the signal processing algorithms. In this special session, we will focus on the recent advances in cryptography and image/video processing. In particular, we would like to investigate the recent developments in image and video encryption, the security and privacy aspects of image processing and efficient cryptographic techniques for image and video processing.
- 2706: William Puech, Zekeriya Erkin, et al., "EMERGING CRYPTOGRAPHIC CHALLENGES IN IMAGE AND VIDEO PROCESSING"
- 1524: Siu-Kei Au Yeung and Bing Zeng, "A NEW DESIGN OF MULTIPLE TRANSFORMS FOR PERCEPTUAL VIDEO ENCRYPTION"
- 2114: Shantanu Rane and Wei Sun, "AN ATTRIBUTE-BASED FRAMEWORK FOR PRIVACY PRESERVING IMAGE QUERYING"
- 2429: Ying Luo, Tommaso Pignata, et al., "AN EFFICIENT PROTOCOL FOR PRIVATE IRIS-CODE MATCHING BY MEANS OF GARBLED CIRCUITS"
- 2916: Juan Ramon Troncoso-Pastoriza and Fernando Perez-Gonzalez, "FULLY HOMOMORPHIC FACES"
- 3242: Thomas Stutz and Andreas Uhl, "COMPLEXITY ANALYSIS OF THE KEY-DEPENDENT WAVELET PACKET TRANSFORM FOR JPEG2000 ENCRYPTION"
- 3297: Loic Dubois, William Purech and Jacques Blanc-Talon, "REDUCED SELECTIVE ENCRYPTION OF INTRA AND INTER FRAMES OF H.264/AVC USING PSYCHOVISUAL METRICS"
- 3472: Yi-Chong Zeng and Miao-Fen Chueh, "ABNORMAL ACTION DETECTION IN ENCRYPTED VIDEO-BASED SURVEILLANCE SYSTEM"