Depth Stream Research Articles

3-D Convolutional Neural Networks for RGB-D Salient Object Detection and Beyond.

RGB-depth (RGB-D) salient object detection (SOD) recently has attracted increasing research interest, and many deep learning methods based on encoder-decoder architectures have emerged. However, most existing RGB-D SOD models conduct explicit and controllable cross-modal feature fusion either in the single encoder or decoder stage, which hardly guarantees sufficient cross-modal fusion ability. To this end, we make the first attempt in addressing RGB-D SOD through 3-D convolutional neural networks. The proposed model, named RD3D, aims at prefusion in the encoder stage and in-depth fusion in the decoder stage to effectively promote the full integration of RGB and depth streams. Specifically, RD3D first conducts prefusion across RGB and depth modalities through a 3-D encoder obtained by inflating 2-D ResNet and later provides in-depth feature fusion by designing a 3-D decoder equipped with rich back-projection paths (RBPPs) for leveraging the extensive aggregation ability of 3-D convolutions. Toward an improved model RD3D+, we propose to disentangle the conventional 3-D convolution into successive spatial and temporal convolutions and, meanwhile, discard unnecessary zero padding. This eventually results in a 2-D convolutional equivalence that facilitates optimization and reduces parameters and computation costs. Thanks to such a progressive-fusion strategy involving both the encoder and the decoder, effective and thorough interactions between the two modalities can be exploited and boost detection accuracy. As an additional boost, we also introduce channel-modality attention and its variant after each path of RBPP to attend to important features. Extensive experiments on seven widely used benchmark datasets demonstrate that RD3D and RD3D+ perform favorably against 14 state-of-the-art RGB-D SOD approaches in terms of five key evaluation metrics. Our code will be made publicly available at https://github.com/PPOLYpubki/RD3D.

CUSCO: An Unobtrusive Custom Secure Audio-Visual Recording System for Ambient Assisted Living.

The ubiquity of digital technology has facilitated detailed recording of human behaviour. Ambient technology has been used to capture behaviours in a broad range of applications ranging from healthcare and monitoring to assessment of cooperative work. However, existing systems often face challenges in terms of autonomy, usability, and privacy. This paper presents a portable, easy-to-use and privacy-preserving system for capturing behavioural signals unobtrusively in home or in office settings. The system focuses on the capture of audio, video, and depth imaging. It is based on a device built on a small-factor platform that incorporates ambient sensors which can be integrated with the audio and depth video hardware for multimodal behaviour tracking. The system can be accessed remotely and integrated into a network of sensors. Data are encrypted in real time to ensure safety and privacy. We illustrate uses of the device in two different settings, namely, a healthy-ageing IoT application, where the device is used in conjunction with a range of IoT sensors to monitor an older person's mental well-being at home, and a healthcare communication quality assessment application, where the device is used to capture a patient-clinician interaction for consultation quality appraisal. CUSCO can automatically detect active speakers, extract acoustic features, record video and depth streams, and recognise emotions and cognitive impairment with promising accuracy.

ECW-EGNet: Exploring Cross-ModalWeighting and edge-guided decoder network for RGB-D salient object detection

Existing RGB-D salient object detection (SOD) techniques concentrate on combining data from multiple modalities (e.g., depth and RGB) and extracting multi-scale data for improved saliency reasoning. However, they frequently perform poorly as a factor of the drawbacks of low-quality depth maps and the lack of correlation between the extracted multi-scale data. In this paper, we propose a Exploring Cross-ModalWeighting and Edge-Guided Decoder Network (ECW-EGNet) for RGB-D SOD, which includes three prominent components. Firstly, we deploy a Cross-Modality Weighting Fusion (CMWF) module that utilizes Channel-Spatial Attention Feature Enhancement (CSAE) mechanism and Depth-Quality Assessment (DQA) mechanism to achieve the cross-modal feature interaction. The former parallels channel attention and spatial attention enhances the features of extracted RGB streams and depth streams while the latter assesses the depth-quality reduces the detrimental influence of the low-quality depth maps during the cross-modal fusion. Then, in order to effectively integrate multi-scale features for high-level and produce salient objects with precise locations, we construct a Bi-directional Scale- Correlation Convolution (BSCC) module in a bi-directional structure. Finally, we construct an Edge-Guided (EG) decoder that uses the edge detection operator to obtain edge masks to guide the enhancement of salient map edge details. The comprehensive experiments on five benchmark RGB-D SOD datasets demonstrate that the proposed ECW-EGNet outperforms 21 state-of-the-art (SOTA) saliency detectors in four widely used evaluation metrics.

CMOT: A cross-modality transformer for RGB-D fusion in person re-identification with online learning capabilities

Person re-identification (reID) is a crucial aspect of intelligent surveillance systems, enabling the recognition of individuals across non-overlapping camera views. As compared to the RGB modality, RGB-D based reID has the potential to achieve robust and high performance by leveraging rich complementary features of both modalities, making it applicable in various occluded scenarios. However, current multimodal reID approaches often rely on late fusion or feature-level fusion techniques to combine multiple modalities, which limits their ability to proficiently exploit complementary visual and depth-related semantic information and capture complex interactions between unimodal features. To address these limitations, this paper introduces a cross-modality online transformer (CMOT) for RGB-D based person reID with online learning capabilities, which effectively utilizes both RGB and depth modalities for the extraction of spatio-temporal features and fuses across modalities. Our CMOT is composed of three main components: (1) a hypothesis generation module based on a person detector and tracker, (2) dual-stream feature extractors via convolutional neural networks (CNNs), (3) and a fusion transformer based on a self-attention-driven self-attentive modality refinement module (SAMR) and a cross-attention-driven cross-attentive modality interaction module (CAMI) to refine and fuse RGB-D complementary features extracted from the dual-stream, RGB and depth stream, CNNs. Additionally, we introduce a bottleneck enhancement feed-forward block to enhance the model’s representation capability within SAMR and CAMI, significantly reducing parameters and computations compared to the traditional feed-forward network. Moreover, we design the triplet loss function with distance measuring ability for incorporating online learning and finally CMOT works as a few-shot network for reID. Experimental results on three RGB-D person re-identification datasets, namely BIWI RGBD-ID, RobotPKU RGBD-ID, and TVPR2, demonstrate the effectiveness and robustness of CMOT.

Depth cue enhancement and guidance network for RGB-D salient object detection

Depth maps have been proven profitable to provide supplements for salient object detection in recent years. However, most RGB-D salient object detection approaches ignore that there are usually low-quality depth maps, which will inevitably result in unsatisfactory results. In this paper, we propose a depth cue enhancement and guidance network (DEGNet) for RGB-D salient object detection by exploring the depth quality enhancement and utilizing the depth cue guidance to generate predictions with highlighted objects and suppressed backgrounds. Specifically, a depth cue enhancement module is designed to generate high-quality depth maps by enhancing the contrast between the foreground and the background. Then considering the different characteristics of unimodal RGB and depth features, we use different feature enhancement strategies to strengthen the representation capability of side-output unimodal features. Moreover, we propose a depth-guided feature fusion module to excavate depth cues provided by the depth stream to guide the fusion of multi-modal features by fully making use of different modal properties, thus generating discriminative cross-modal features. Besides, we aggregate cross-modal features at different levels to obtain the final prediction by adopting a pyramid feature shrinking structure. Experimental results on six benchmark datasets demonstrate that the proposed network DEGNet outperforms 17 state-of-the-art methods.

Evaluation of Hand Washing Procedure Using Vision-Based Frame Level and Spatio-Temporal Level Data Models

Hand hygiene is obligatory for all healthcare workers and vital for patient care. During COVID-19, adequate hand washing was among recommended measures for preventing virus transmission. A general hand-washing procedure consisting several steps is recommended by World Health Organization for ensuring hand hygiene. This process can vary from person to person and human supervision for inspection would be impractical. In this study, we propose computer vision-based new methods using 12 different neural network models and 4 different data models (RGB, Point Cloud, Point Gesture Map, Projection) for the classification of 8 universally accepted hand-washing steps. These methods can also perform well under situations where the order of steps is not observed or the duration of steps are varied. Using a custom dataset, we achieved 100% accuracy with one of the models, and 94.23% average accuracy for all models. We also developed a real-time robust data acquisition technique where RGB and depth streams from Kinect 2.0 camera were utilized. Results showed that with the proposed methods and data models, efficient hand hygiene control is possible.

Light field angular super-resolution based on intrinsic and geometric information

Light field imaging can encode abundant scene information, including the intensity and direction of light rays, into 4D light field images. However, the limited number of sensors in commercial light field cameras leads to a trade-off between spatial and angular resolutions. In this paper, an angular super-resolution framework is proposed to synthesize new views and overcome hardware restrictions. First, light field intrinsic feature convolution is proposed to extract intrinsic information, i.e., scene content, complete view correlations, and epipolar structures. Consequently, spatial, angular, and cross-domain information can be preserved in the extracted features. Second, the spatial–angular and depth streams are proposed based on the light field intrinsic feature convolution to synthesize high angular resolution light fields. The spatial–angular stream utilizes the light field intrinsic information to improve the angular resolution, whereas the depth stream disentangles the geometric information from the extracted light field intrinsic features, which is used to warp the given sub-aperture images to the new view positions. Both streams can synthesize high-quality intermediate results, where the intrinsic and geometric information are utilized separately. Finally, a confidence-based stream fusion module is proposed to fuse the outputs from the two streams, achieving the joint employment of light field intrinsic and geometric information and solving the problem of insufficient information exploration in the current methods. We conduct a series of experiments to validate the effectiveness of each component in the framework and demonstrate that our method can achieve state-of-the-art performance in various scenes.

FASFLNet: feature adaptive selection and fusion lightweight network for RGB-D indoor scene parsing.

RGB-D indoor scene parsing is a challenging task in computer vision. Conventional scene-parsing approaches based on manual feature extraction have proved inadequate in this area because indoor scenes are both unordered and complex. This study proposes a feature adaptive selection, and fusion lightweight network (FASFLNet) for RGB-D indoor scene parsing that is both efficient and accurate. The proposed FASFLNet utilizes a lightweight classification network (MobileNetV2), constituting the backbone of the feature extraction. This lightweight backbone model guarantees that FASFLNet is not only highly efficient but also provides good performance in terms of feature extraction. The additional information provided by depth images (specifically, spatial information such as the shape and scale of objects) is used in FASFLNet as supplemental information for feature-level adaptive fusion between the RGB and depth streams. Furthermore, during decoding, the features of different layers are fused from top-bottom and integrated at different layers for final pixel-level classification, resulting in an effect similar to that of pyramid supervision. Experimental results obtained on the NYU V2 and SUN RGB-D datasets indicate that the proposed FASFLNet outperforms existing state-of-the-art models and is both highly efficient and accurate.

SSEC: Semantic Segmentation and Ensemble Classification Framework for Static Hand Gesture Recognition using RGB-D Data

Hand Gesture Recognition (HGR) refers to identifying various hand postures used in Sign Language Recognition (SLR) and Human Computer Interaction (HCI) applications. Complex background in uncontrolled environmental condition is the major challenging issue which impacts the recognition accuracy of HGR system. This can be effectively addressed by discarding the background using suitable semantic segmentation method, where it predicts the hand region pixels into foreground and rest of the pixels into background. In this paper, we have analyzed and evaluated well known semantic segmentation architectures for hand region segmentation using both RGB and depth data. Further, ensemble of segmented RGB and depth stream is used for hand gesture classification through probability score fusion. Experimental results shows that the proposed novel framework of Semantic Segmentation and Ensemble Classification (SSEC) is suitable for static hand gesture recognition and achieved F1-score of 88.91% on OUHANDS test dataset.

Encoder deep interleaved network with multi-scale aggregation for RGB-D salient object detection

Recently, RGB-D salient object detection (SOD) has aroused widespread research interest. Existing RGB-D SOD approaches mainly consider the cross-modal information fusion in the decoder. And their multi-modal interaction mainly concentrates on the same level of features between RGB stream and depth stream. They do not deeply explore the coherence of multi-model features at different levels. In this paper, we design a two-stream deep interleaved encoder network to extract RGB and depth information and realize their mixing simultaneously. This network allows us to gradually learn multi-modal representation at different levels from shallow to deep. Moreover, to further fuse multi-modal features in the decoding stage, we propose a cross-modal mutual guidance module and a residual multi-scale aggregation module to implement the global guidance and local refinement of the salient region. Extensive experiments on six benchmark datasets demonstrate that the proposed approach performs favorably against most state-of-the-art methods under different evaluation metrics. During the testing stage, this model can run at a real-time speed of 93 FPS.

When CNNs meet random RNNs: Towards multi-level analysis for RGB-D object and scene recognition

Recognizing objects and scenes are two challenging but essential tasks in image understanding. In particular, the use of RGB-D sensors in handling these tasks has emerged as an important area of focus for better visual understanding. Meanwhile, deep neural networks, specifically convolutional neural networks (CNNs), have become widespread and have been applied to many visual tasks by replacing hand-crafted features with effective deep features. However, it is an open problem how to exploit deep features from a multi-layer CNN model effectively. In this paper, we propose a novel two-stage framework that extracts discriminative feature representations from multi-modal RGB-D images for object and scene recognition tasks. In the first stage, a pretrained CNN model has been employed as a backbone to extract visual features at multiple levels. The second stage maps these features into high level representations with a fully randomized structure of recursive neural networks (RNNs) efficiently. To cope with the high dimensionality of CNN activations, a random weighted pooling scheme has been proposed by extending the idea of randomness in RNNs. Multi-modal fusion has been performed through a soft voting approach by computing weights based on individual recognition confidences (i.e. SVM scores) of RGB and depth streams separately. This produces consistent class label estimation in final RGB-D classification performance. Extensive experiments verify that fully randomized structure in RNN stage encodes CNN activations to discriminative solid features successfully. Comparative experimental results on the popular Washington RGB-D Object and SUN RGB-D Scene datasets show that the proposed approach achieves superior or on-par performance compared to state-of-the-art methods both in object and scene recognition tasks. Code is available at https://github.com/acaglayan/CNN_randRNN.

DMRA: Depth-Induced Multi-Scale Recurrent Attention Network for RGB-D Saliency Detection.

In this work, we propose a novel depth-induced multi-scale recurrent attention network for RGB-D saliency detection, named as DMRA. It achieves dramatic performance especially in complex scenarios. There are four main contributions of our network that are experimentally demonstrated to have significant practical merits. First, we design an effective depth refinement block using residual connections to fully extract and fuse cross-modal complementary cues from RGB and depth streams. Second, depth cues with abundant spatial information are innovatively combined with multi-scale contextual features for accurately locating salient objects. Third, a novel recurrent attention module inspired by Internal Generative Mechanism of human brain is designed to generate more accurate saliency results via comprehensively learning the internal semantic relation of the fused feature and progressively optimizing local details with memory-oriented scene understanding. Finally, a cascaded hierarchical feature fusion strategy is designed to promote efficient information interaction of multi-level contextual features and further improve the contextual representability of model. In addition, we introduce a new real-life RGB-D saliency dataset containing a variety of complex scenarios that has been widely used as a benchmark dataset in recent RGB-D saliency detection research. Extensive empirical experiments demonstrate that our method can accurately identify salient objects and achieve appealing performance against 18 state-of-the-art RGB-D saliency models on nine benchmark datasets.

Single Depth View Based Real-Time Reconstruction of Hand-Object Interactions

Reconstructing hand-object interactions is a challenging task due to strong occlusions and complex motions. This article proposes a real-time system that uses a single depth stream to simultaneously reconstruct hand poses, object shape, and rigid/non-rigid motions. To achieve this, we first train a joint learning network to segment the hand and object in a depth image, and to predict the 3D keypoints of the hand. With most layers shared by the two tasks, computation cost is saved for the real-time performance. A hybrid dataset is constructed here to train the network with real data (to learn real-world distributions) and synthetic data (to cover variations of objects, motions, and viewpoints). Next, the depth of the two targets and the keypoints are used in a uniform optimization to reconstruct the interacting motions. Benefitting from a novel tangential contact constraint, the system not only solves the remaining ambiguities but also keeps the real-time performance. Experiments show that our system handles different hand and object shapes, various interactive motions, and moving cameras.

Sign language recognition using Kinect sensor based on color stream and skeleton points

This paper presents a sign language recognition system based on color stream and skeleton points. Several approaches have been established to address sign language recognition problems. However, most of the previous approaches still have poor recognition accuracy. The proposed approach uses Kinect sensor based on color stream and skeleton points from the depth stream to improved recognition accuracy. Techniques within this approach use hand trajectories and hand shapes in combating sign recognition challenges. Therefore, for a particular sign a representative feature vector is extracted, which consists of hand trajectories and hand shapes. A sparse dictionary learning algorithm, Label Consistent K-SVD (LC-KSVD) is applied to obtain a discriminative dictionary. Based on that, the system was further developed to a new classification approach for better results. The proposed system was fairly evaluated based on 21 sign words including one-handed signs and two-handed signs. It was observed that the proposed system gets high recognition accuracy of 98.25%, and obtained an average accuracy of 95.34% for signer independent recognition.
 Keywords: Sign language, Color stream, Skeleton points, Kinect sensor, Discriminative dictionary.

RGB-D Salient Object Detection via 3D Convolutional Neural Networks

RGB-D salient object detection (SOD) recently has attracted increasing research interest and many deep learning methods based on encoder-decoder architectures have emerged. However, most existing RGB-D SOD models conduct feature fusion either in the single encoder or the decoder stage, which hardly guarantees sufficient cross-modal fusion ability. In this paper, we make the first attempt in addressing RGB-D SOD through 3D convolutional neural networks. The proposed model, named RD3D, aims at pre-fusion in the encoder stage and in-depth fusion in the decoder stage to effectively promote the full integration of RGB and depth streams. Specifically, RD3D first conducts pre-fusion across RGB and depth modalities through an inflated 3D encoder, and later provides in-depth feature fusion by designing a 3D decoder equipped with rich back-projection paths (RBPP) for leveraging the extensive aggregation ability of 3D convolutions. With such a progressive fusion strategy involving both the encoder and decoder, effective and thorough interaction between the two modalities can be exploited and boost the detection accuracy. Extensive experiments on six widely used benchmark datasets demonstrate that RD3D performs favorably against 14 state-of-the-art RGB-D SOD approaches in terms of four key evaluation metrics. Our code will be made publicly available: https://github.com/PPOLYpubki/RD3D.

CDNet: Complementary Depth Network for RGB-D Salient Object Detection.

Current RGB-D salient object detection (SOD) methods utilize the depth stream as complementary information to the RGB stream. However, the depth maps are usually of low-quality in existing RGB-D SOD datasets. Most RGB-D SOD networks trained with these datasets would produce error-prone results. In this paper, we propose a novel Complementary Depth Network (CDNet) to well exploit saliency-informative depth features for RGB-D SOD. To alleviate the influence of low-quality depth maps to RGB-D SOD, we propose to select saliency-informative depth maps as the training targets and leverage RGB features to estimate meaningful depth maps. Besides, to learn robust depth features for accurate prediction, we propose a new dynamic scheme to fuse the depth features extracted from the original and estimated depth maps with adaptive weights. What's more, we design a two-stage cross-modal feature fusion scheme to well integrate the depth features with the RGB ones, further improving the performance of our CDNet on RGB-D SOD. Experiments on seven benchmark datasets demonstrate that our CDNet outperforms state-of-the-art RGB-D SOD methods. The code is publicly available at https://github.com/blanclist/CDNet.

Joint learning of convolution neural networks for RGB‐D‐based human action recognition

RGB-D-based human action recognition aims to learn distinctive features from different modalities and has shown good progress in practice. However, it is difficult to improve the recognition performance through directly training multiple individual convolutional networks (ConvNets) and fusing features later because complmentary information between different modalities cannot be learned. To address this issue, this Letter proposes a single two-stream ConvNets framework for multimodality learning that extract features through RGB and depth streams. Specifically, the authors first represent RGB-D sequence to motion images as the inputs of the proposed ConvNets for obtaining spatial–temporal information. Then, a features fusion and joint training strategy is adapted to learn RGB-D complementary features simultaneously. Experimental results on benchmark NTU RGB+D 120 dataset validate the effectiveness of the proposed framework and demonstrate that two-stream ConvNets outperforms the current state-of-the-art approaches.

Age‐0 Smallmouth Bass abundance depends on physicochemical conditions and stream network position

AbstractStream fish survival and recruitment are products of a physicochemical environment that affects growth and provides refuge; yet, the drivers of spatiotemporal variation in juvenile fish abundance remain unclear. Understanding how physicochemical conditions drive spatial and temporal patterns in fish abundances provides insight into how conditions across stream networks influence fish population success, thereby providing direction to managers about the types and locations of conservation actions that would be most beneficial. Using snorkel and habitat surveys of 120 sites sampled from 2015 to 2017, we evaluated the multiscale relationships among physicochemical features, hydrology, and age‐0 Smallmouth Bass (Micropterus dolomieu velox) abundance in relation to network spatial position. Abundance of age‐0 bass was spatiotemporally variable in relation to a July streamflow–network position interaction, a pool depth–stream size interaction, and a stream temperature–network position interaction. High flows at the end of the nesting season were related to lower age‐0 abundance, but this effect was dampened in stream reaches in close proximity to larger mainstems. In small streams, reaches with deeper pool habitat supported higher age‐0 bass abundances, but this trend was not apparent in larger tributaries and mainstem systems. Generally, colder streams had lower age‐0 Smallmouth Bass abundance, though this relationship was not apparent in reaches adjacent to larger streams that generally supported higher age‐0 bass abundances. Conservation actions that (1) facilitate habitat connectivity within and among streams, (2) limit future anthropogenic practices that alter natural geomorphology by creating shallower stream channels, and (3) maintain adequate flow magnitude and timing to support channel complexity (e.g., deeper pools within smaller catchments) would be most beneficial to supporting rearing habitat for age‐0 riverine Smallmouth Bass.

TSNet: Three-Stream Self-Attention Network for RGB-D Indoor Semantic Segmentation

This article proposes a three-stream self-attention network (TSNet) for indoor semantic segmentation comprising two asymmetric input streams (asymmetric encoder structure) and a cross-modal distillation stream with a self-attention module. The two asymmetric input streams are ResNet34 for the red-green-blue (RGB) stream and VGGNet16 for the depth stream. Accompanying the RGB and depth streams, a cross-modal distillation stream with a self-attention module extracts new RGB plus depth features in each level in the bottom-up path. In addition, while using bilinear upsampling to recover the spatial resolution of the feature map, we incorporated the feature information of both the RGB flow and the depth flow through the self-attention module. We constructed the NYU Depth V2 dataset to evaluate the TSNet and achieved results comparable to those of current state-of-the-art methods.

Multi-level progressive parallel attention guided salient object detection for RGB-D images

Detecting salient objects in RGB-D images attracts more and more attention in recent years. It benefits from the widespread use of depth sensors and can be applied in the comprehensive understanding of RGB-D images. Existing models focus on double-stream networks which transfer from color stream to depth stream, but depth stream with one channel information cannot learn the same feature as color stream with three channels information even if HHA representation is adopted. In our works, RGB-D four-channels input is chosen, and meanwhile, progressive parallel spatial and channel attention mechanisms are performed to improve feature representation. Spatial and channel attention can pay more attention on partial positions and channels in the image which show higher response to salient objects. Both attentive features are optimized by attentive feature from higher layer, respectively, and parallel fed into recurrent convolutional layer to generate side-output saliency maps guided by saliency map from higher layer. Last multi-level saliency maps are fused together from multi-scale perspective. Experiments on benchmark datasets demonstrate that parallel attention mechanism and progressive optimization operation play an important role in improving the accuracy of salient object detection, and our model outperforms state-of-the-art models in evaluation matrices.