Deep state-space models (DSSMs) have gained popularity in recent years due to their potent modeling capacity for dynamic systems. However, existing DSSM works are limited to single-task modeling, which requires retraining with historical task data upon revisiting a forepassed task. To address this limitation, we propose continual learning DSSMs (CLDSSMs), which are capable of adapting to evolving tasks without catastrophic forgetting. Our proposed CLDSSMs integrate mainstream regularization-based continual learning (CL) methods, ensuring efficient updates with constant computational and memory costs for modeling multiple dynamic systems. We also conduct a comprehensive cost analysis of each CL method applied to the respective CLDSSMs, and demonstrate the efficacy of CLDSSMs through experiments on real-world datasets. The results corroborate that while various competing CL methods exhibit different merits, the proposed CLDSSMs consistently outperform traditional DSSMs in terms of effectively addressing catastrophic forgetting, enabling swift and accurate parameter transfer to new tasks.
Multi-Agent Combinatorial Path Finding (MCPF) seeks collision-free paths for multiple agents from their initial locations to destinations, visiting a set of intermediate target locations in the middle of the paths, while minimizing the sum of arrival times. While a few approaches have been developed to handle MCPF, most of them simply direct the agent to visit the targets without considering the task duration, i.e., the amount of time needed for an agent to execute the task (such as picking an item) at a target location. MCPF is NP-hard to solve to optimality, and the inclusion of task duration further complicates the problem. This paper investigates heterogeneous task duration, where the duration can be different with respect to both the agents and targets. We develop two methods, where the first method post-processes the paths planned by any MCPF planner to include the task duration and has no solution optimality guarantee; and the second method considers task duration during planning and is able to ensure solution optimality. The numerical and simulation results show that our methods can handle up to 20 agents and 50 targets in the presence of task duration, and can execute the paths subject to robot motion disturbance.
Path planning plays a crucial role in various autonomy applications, and RRT* is one of the leading solutions in this field. In this paper, we propose the utilization of vertex-based networks to enhance the sampling process of RRT*, leading to more efficient path planning. Our approach focuses on critical vertices along the optimal paths, which provide essential yet sparser abstractions of the paths. We employ focal loss to address the associated data imbalance issue, and explore different masking configurations to determine practical tradeoffs in system performance. Through experiments conducted on randomly generated floor maps, our solutions demonstrate significant speed improvements, achieving over a 400% enhancement compared to the baseline model.
This report presents a brief description of our winning solution to the AVA Active Speaker Detection (ASD) task at ActivityNet Challenge 2022. Our underlying model UniCon+ continues to build on our previous work, the Unified Context Network (UniCon) and Extended UniCon which are designed for robust scene-level ASD. We augment the architecture with a simple GRU-based module that allows information of recurring identities to flow across scenes through read and update operations. We report a best result of 94.47% mAP on the AVA-ActiveSpeaker test set, which continues to rank first on this year's challenge leaderboard and significantly pushes the state-of-the-art.
Depth Estimation and Object Detection Recognition play an important role in autonomous driving technology under the guidance of deep learning artificial intelligence. We propose a hybrid structure called RealNet: a co-design method combining the model-streamlined recognition algorithm, the depth estimation algorithm with information fusion, and deploying them on the Jetson-Nano for unmanned vehicles with monocular vision sensors. We use ROS for experiment. The method proposed in this paper is suitable for mobile platforms with high real-time request. Innovation of our method is using information fusion to compensate the problem of insufficient frame rate of output image, and improve the robustness of target detection and depth estimation under monocular vision.Object Detection is based on YOLO-v5. We have simplified the network structure of its DarkNet53 and realized a prediction speed up to 0.01s. Depth Estimation is based on the VNL Depth Estimation, which considers multiple geometric constraints in 3D global space. It calculates the loss function by calculating the deviation of the virtual normal vector VN and the label, which can obtain deeper depth information. We use PnP fusion algorithm to solve the problem of insufficient frame rate of depth map output. It solves the motion estimation depth from three-dimensional target to two-dimensional point based on corner feature matching, which is faster than VNL calculation. We interpolate VNL output and PnP output to achieve information fusion. Experiments show that this can effectively eliminate the jitter of depth information and improve robustness. At the control end, this method combines the results of target detection and depth estimation to calculate the target position, and uses a pure tracking control algorithm to track it.
We introduce a new efficient framework, the Unified Context Network (UniCon), for robust active speaker detection (ASD). Traditional methods for ASD usually operate on each candidate's pre-cropped face track separately and do not sufficiently consider the relationships among the candidates. This potentially limits performance, especially in challenging scenarios with low-resolution faces, multiple candidates, etc. Our solution is a novel, unified framework that focuses on jointly modeling multiple types of contextual information: spatial context to indicate the position and scale of each candidate's face, relational context to capture the visual relationships among the candidates and contrast audio-visual affinities with each other, and temporal context to aggregate long-term information and smooth out local uncertainties. Based on such information, our model optimizes all candidates in a unified process for robust and reliable ASD. A thorough ablation study is performed on several challenging ASD benchmarks under different settings. In particular, our method outperforms the state-of-the-art by a large margin of about 15% mean Average Precision (mAP) absolute on two challenging subsets: one with three candidate speakers, and the other with faces smaller than 64 pixels. Together, our UniCon achieves 92.0% mAP on the AVA-ActiveSpeaker validation set, surpassing 90% for the first time on this challenging dataset at the time of submission. Project website: https://unicon-asd.github.io/.
Lip reading is the task of recognizing the speech content by analyzing movements in the lip region when people are speaking. Observing on the continuity in adjacent frames in the speaking process, and the consistency of the motion patterns among different speakers when they pronounce the same phoneme, we model the lip movements in the speaking process as a sequence of apparent deformations in the lip region. Specifically, we introduce a Deformation Flow Network (DFN) to learn the deformation flow between adjacent frames, which directly captures the motion information within the lip region. The learned deformation flow is then combined with the original grayscale frames with a two-stream network to perform lip reading. Different from previous two-stream networks, we make the two streams learn from each other in the learning process by introducing a bidirectional knowledge distillation loss to train the two branches jointly. Owing to the complementary cues provided by different branches, the two-stream network shows a substantial improvement over using either single branch. A thorough experimental evaluation on two large-scale lip reading benchmarks is presented with detailed analysis. The results accord with our motivation, and show that our method achieves state-of-the-art or comparable performance on these two challenging datasets.
Recent advances in deep learning have heightened interest among researchers in the field of visual speech recognition (VSR). Currently, most existing methods equate VSR with automatic lip reading, which attempts to recognise speech by analysing lip motion. However, human experience and psychological studies suggest that we do not always fix our gaze at each other's lips during a face-to-face conversation, but rather scan the whole face repetitively. This inspires us to revisit a fundamental yet somehow overlooked problem: can VSR models benefit from reading extraoral facial regions, i.e. beyond the lips? In this paper, we perform a comprehensive study to evaluate the effects of different facial regions with state-of-the-art VSR models, including the mouth, the whole face, the upper face, and even the cheeks. Experiments are conducted on both word-level and sentence-level benchmarks with different characteristics. We find that despite the complex variations of the data, incorporating information from extraoral facial regions, even the upper face, consistently benefits VSR performance. Furthermore, we introduce a simple yet effective method based on Cutout to learn more discriminative features for face-based VSR, hoping to maximise the utility of information encoded in different facial regions. Our experiments show obvious improvements over existing state-of-the-art methods that use only the lip region as inputs, a result we believe would probably provide the VSR community with some new and exciting insights.
Large-scale datasets have successively proven their fundamental importance in several research fields, especially for early progress in some emerging topics. In this paper, we focus on the problem of visual speech recognition, also known as lipreading, which has received an increasing interest in recent years. We present a naturally-distributed large-scale benchmark for lip reading in the wild, named LRW-1000, which contains 1000 classes with about 745,187 samples from more than 2000 individual speakers. Each class corresponds to the syllables of a Mandarin word which is composed of one or several Chinese characters. To the best of our knowledge, it is the largest word-level lipreading dataset and also the only public large-scale Mandarin lipreading dataset. This dataset aims at covering a "natural" variability over different speech modes and imaging conditions to incorporate challenges encountered in practical applications. This benchmark shows a large variation over several aspects, including the number of samples in each class, resolution of videos, lighting conditions, and speakers' attributes such as pose, age, gender, and make-up. Besides a detailed description of the dataset and its collection pipeline, we evaluate the popular lipreading methods and perform a thorough analysis of the results from several aspects. The results demonstrate the consistency and challenges of our dataset, which may open up some new promising directions for future work. The dataset and corresponding codes will be public for academic research use.
Visual speech recognition is the task to decode the speech content from a video based on visual information, especially the movements of lips. It is also referenced as lipreading. Motivated by two problems existing in lipreading, words with similar pronunciation and the variation of word duration, we propose a novel 3D Feature Pyramid Attention (3D-FPA) module to jointly improve the representation power of features in both the spatial and temporal domains. Specifically, the input features are downsampled for 3 times in both the spatial and temporal dimensions to construct spatiotemporal feature pyramids. Then high-level features are upsampled and combined with low-level features, finally generating a pixel-level soft attention mask to be multiplied with the input features.It enhances the discriminative power of features and exploits the temporal multi-scale information while decoding the visual speeches. Also, this module provides a new method to construct and utilize temporal pyramid structures in video analysis tasks. The field of temporal featrue pyramids are still under exploring compared to the plentiful works on spatial feature pyramids for image analysis tasks. To validate the effectiveness and adaptability of our proposed module, we embed the module in a sentence-level lipreading model, LipNet, with the result of 3.6% absolute decrease in word error rate, and a word-level model, with the result of 1.4% absolute improvement in accuracy.