PhD student, City University of Hong Kong
I track objects appeared in the videos.
Visual tracking aims to predict target object locations in the videos. The input is the location of target object in the first frame and we propose methods to predict its locations in the following frames.
Abstract: The problem of determining whether an object is in motion, irrespective of the camera motion, is far from being solved. We address this challenging task by learning motion patterns in videos. The core of our approach is a fully convolutional network, which is learnt entirely from synthetic video sequences, and their ground-truth optical flow and motion segmentation. This encoder-decoder style architecture first learns a coarse representation of the optical flow field features, and then refines it iteratively to produce motion labels at the original high-resolution. The output label of each pixel denotes whether it has undergone independent motion, i.e., irrespective of the camera motion. We demonstrate the benefits of this learning framework on the moving object segmentation task, where the goal is to segment all the objects in motion. To this end we integrate an objectness measure into the framework. Our approach outperforms the top method on the recently released DAVIS benchmark dataset, comprising real-world sequences, by 5.6%. We also evaluate on the Berkeley motion segmentation database, achieving state-of-the-art results.
Pub.: 21 Dec '16, Pinned: 24 Aug '17
Abstract: Models based on deep convolutional networks have dominated recent image interpretation tasks; we investigate whether models which are also recurrent, or "temporally deep", are effective for tasks involving sequences, visual and otherwise. We develop a novel recurrent convolutional architecture suitable for large-scale visual learning which is end-to-end trainable, and demonstrate the value of these models on benchmark video recognition tasks, image description and retrieval problems, and video narration challenges. In contrast to current models which assume a fixed spatio-temporal receptive field or simple temporal averaging for sequential processing, recurrent convolutional models are "doubly deep"' in that they can be compositional in spatial and temporal "layers". Such models may have advantages when target concepts are complex and/or training data are limited. Learning long-term dependencies is possible when nonlinearities are incorporated into the network state updates. Long-term RNN models are appealing in that they directly can map variable-length inputs (e.g., video frames) to variable length outputs (e.g., natural language text) and can model complex temporal dynamics; yet they can be optimized with backpropagation. Our recurrent long-term models are directly connected to modern visual convnet models and can be jointly trained to simultaneously learn temporal dynamics and convolutional perceptual representations. Our results show such models have distinct advantages over state-of-the-art models for recognition or generation which are separately defined and/or optimized.
Pub.: 31 May '16, Pinned: 24 Aug '17
Abstract: Convolutional neural network (CNN) models have achieved tremendous success in many visual detection and recognition tasks. Unfortunately, visual tracking, a fundamental computer vision problem, is not handled well using the existing CNN models, because most object trackers implemented with CNN do not effectively leverage temporal and contextual information among consecutive frames. Recurrent neural network (RNN) models, on the other hand, are often used to process text and voice data due to their ability to learn intrinsic representations of sequential and temporal data. Here, we propose a novel neural network tracking model that is capable of integrating information over time and tracking a selected target in video. It comprises three components: a CNN extracting best tracking features in each video frame, an RNN constructing video memory state, and a reinforcement learning (RL) agent making target location decisions. The tracking problem is formulated as a decision-making process, and our model can be trained with RL algorithms to learn good tracking policies that pay attention to continuous, inter-frame correlation and maximize tracking performance in the long run. We compare our model with an existing neural-network based tracking method and show that the proposed tracking approach works well in various scenarios by performing rigorous validation experiments on artificial video sequences with ground truth. To the best of our knowledge, our tracker is the first neural-network tracker that combines convolutional and recurrent networks with RL algorithms.
Pub.: 31 Jan '17, Pinned: 24 Aug '17