Tracking a region-of-interest (ROI) in a video is still a challenging task. Various high level applications rely on tracking. e.g, motion picture indexing, object recognition, video surveillance, audiovisual postproduction etc. Initially ROI is defined in a reference frame and the purpose is to determine the ROI in subsequent target frames in video sequences. The region was detected by determining the similarity measures between the reference and the target frames. Similarity measures between the frames are determined using two classical methods like sum of squared differences(SSD) and sum of absolute differences(SAD). This paper deals with the method of ROI tracking in video sequences by estimating the colour and geometric features between the frames and the similarity measures was determined using the Kullback-Leibler Divergence. The increase of description features improves the accuracy. Their combination leads to high dimensional PDFs. Tracking experiments were performed on several standard video sequences and its efficiency was proved.

To monitor the drowsiness of driver, this paper describes an efficient method by using three well defined phases. The threephases are facial features detection using Viola Jones, the eyetracking and yawning detection. Once the face is detected, the system is made illumination invariant by segmenting the skin part alone and considering only the chromatic components to reject most of the non face image backgrounds based on skin color. The tracking of eyes and yawning detection are done by correlation coefficient template matching. They can easily capture the image of the personby using a single camera in all directions. The feature vectors from each of the above phases are concatenated and a binarylinear support vector machine classifier is used to classify the consecutive frames into fatigue and nonfatigue states and sound an alarm for the former, if it is above the threshold time. Extensive real time experiments prove that the proposed method is highly efficient in finding the drowsiness and alerting the driver.

Under rainy conditions the impact of rain streaks on images and video is often undesirable. The effects of rain can also severely affect the performance of outdoor vision system. The quality of the image is degraded by rain streaks. Hence it is necessary to remove rain streaks from single image which is a challenging problem. Towards fixing this problem the deep decomposition-composition network is proposed. This paper designs a novel multi-task leaning architecture in an end-to-end manner to reduce the mapping range from input to output and boost the performance. Concretely, a decomposition net is built to split rain images into clean background and rain layers. Different from previous architectures, this model consists of, besides a component representing the desired clean image, an extra component for the rain layer. During the training phase, further employ a composition structure to reproduce the input by the separated clean image and rain information for improving the quality of decomposition. Furthermore, this design is also applicable to other layer decomposition tasks like dust removal. More importantly, this method only requires about 50ms, significantly faster than the competitors, to process a testing image in VGA resolution on a GTX 1080 GPU, making it attractive for practical use