Domain shift degrades the performance of object detection models in practical applications. To alleviate the influence of domain shift, plenty of previous work try to decouple and learn the domain-invariant (common) features from source domains via domain adversarial learning (DAL). However, inspired by causal mechanisms, we find that previous methods ignore the implicit insignificant non-causal factors hidden in the common features. This is mainly due to the single-view nature of DAL. In this work, we present an idea to remove non-causal factors from common features by multi-view adversarial training on source domains, because we observe that such insignificant non-causal factors may still be significant in other latent spaces (views) due to the multi-mode structure of data. To summarize, we propose a Multi-view Adversarial Discriminator (MAD) based domain generalization model, consisting of a Spurious Correlations Generator (SCG) that increases the diversity of source domain by random augmentation and a Multi-View Domain Classifier (MVDC) that maps features to multiple latent spaces, such that the non-causal factors are removed and the domain-invariant features are purified. Extensive experiments on six benchmarks show our MAD obtains state-of-the-art performance.
In this paper, we present an Adaptive Ensemble Learning framework that aims to boost the performance of deep neural networks by intelligently fusing features through ensemble learning techniques. The proposed framework integrates ensemble learning strategies with deep learning architectures to create a more robust and adaptable model capable of handling complex tasks across various domains. By leveraging intelligent feature fusion methods, the Adaptive Ensemble Learning framework generates more discriminative and effective feature representations, leading to improved model performance and generalization capabilities. We conducted extensive experiments and evaluations on several benchmark datasets, including image classification, object detection, natural language processing, and graph-based learning tasks. The results demonstrate that the proposed framework consistently outperforms baseline models and traditional feature fusion techniques, highlighting its effectiveness in enhancing deep learning models' performance. Furthermore, we provide insights into the impact of intelligent feature fusion on model performance and discuss the potential applications of the Adaptive Ensemble Learning framework in real-world scenarios. The paper also explores the design and implementation of adaptive ensemble models, ensemble training strategies, and meta-learning techniques, which contribute to the framework's versatility and adaptability. In conclusion, the Adaptive Ensemble Learning framework represents a significant advancement in the field of feature fusion and ensemble learning for deep neural networks, with the potential to transform a wide range of applications across multiple domains.
We introduce Multi-Source 3D (MS3D), a new self-training pipeline for unsupervised domain adaptation in 3D object detection. Despite the remarkable accuracy of 3D detectors, they often overfit to specific domain biases, leading to suboptimal performance in various sensor setups and environments. Existing methods typically focus on adapting a single detector to the target domain, overlooking the fact that different detectors possess distinct expertise on different unseen domains. MS3D leverages this by combining different pre-trained detectors from multiple source domains and incorporating temporal information to produce high-quality pseudo-labels for fine-tuning. Our proposed Kernel-Density Estimation (KDE) Box Fusion method fuses box proposals from multiple domains to obtain pseudo-labels that surpass the performance of the best source domain detectors. MS3D exhibits greater robustness to domain shift and produces accurate pseudo-labels over greater distances, making it well-suited for high-to-low beam domain adaptation and vice versa. Our method achieved state-of-the-art performance on all evaluated datasets, and we demonstrate that the pre-trained detector's source dataset has minimal impact on the fine-tuned result, making MS3D suitable for real-world applications.
The Structure from Motion (SfM) challenge in computer vision is the process of recovering the 3D structure of a scene from a series of projective measurements that are calculated from a collection of 2D images, taken from different perspectives. SfM consists of three main steps; feature detection and matching, camera motion estimation, and recovery of 3D structure from estimated intrinsic and extrinsic parameters and features. A problem encountered in SfM is that scenes lacking texture or with repetitive features can cause erroneous feature matching between frames. Semantic segmentation offers a route to validate and correct SfM models by labelling pixels in the input images with the use of a deep convolutional neural network. The semantic and geometric properties associated with classes in the scene can be taken advantage of to apply prior constraints to each class of object. The SfM pipeline COLMAP and semantic segmentation pipeline DeepLab were used. This, along with planar reconstruction of the dense model, were used to determine erroneous points that may be occluded from the calculated camera position, given the semantic label, and thus prior constraint of the reconstructed plane. Herein, semantic segmentation is integrated into SfM to apply priors on the 3D point cloud, given the object detection in the 2D input images. Additionally, the semantic labels of matched keypoints are compared and inconsistent semantically labelled points discarded. Furthermore, semantic labels on input images are used for the removal of objects associated with motion in the output SfM models. The proposed approach is evaluated on a data-set of 1102 images of a repetitive architecture scene. This project offers a novel method for improved validation of 3D SfM models.
Regular object detection methods output rectangle bounding boxes, which are unable to accurately describe the actual object shapes. Instance segmentation methods output pixel-level labels, which are computationally expensive for real-time applications. Therefore, a polygon representation is needed to achieve precise shape alignment, while retaining low computation cost. We develop a novel Deformable Polar Polygon Object Detection method (DPPD) to detect objects in polygon shapes. In particular, our network predicts, for each object, a sparse set of flexible vertices to construct the polygon, where each vertex is represented by a pair of angle and distance in the Polar coordinate system. To enable training, both ground truth and predicted polygons are densely resampled to have the same number of vertices with equal-spaced raypoints. The resampling operation is fully differentable, allowing gradient back-propagation. Sparse polygon predicton ensures high-speed runtime inference while dense resampling allows the network to learn object shapes with high precision. The polygon detection head is established on top of an anchor-free and NMS-free network architecture. DPPD has been demonstrated successfully in various object detection tasks for autonomous driving such as traffic-sign, crosswalk, vehicle and pedestrian objects.
Vision-based Transformer have shown huge application in the perception module of autonomous driving in terms of predicting accurate 3D bounding boxes, owing to their strong capability in modeling long-range dependencies between the visual features. However Transformers, initially designed for language models, have mostly focused on the performance accuracy, and not so much on the inference-time budget. For a safety critical system like autonomous driving, real-time inference at the on-board compute is an absolute necessity. This keeps our object detection algorithm under a very tight run-time budget. In this paper, we evaluated a variety of strategies to optimize on the inference-time of vision transformers based object detection methods keeping a close-watch on any performance variations. Our chosen metric for these strategies is accuracy-runtime joint optimization. Moreover, for actual inference-time analysis we profile our strategies with float32 and float16 precision with TensorRT module. This is the most common format used by the industry for deployment of their Machine Learning networks on the edge devices. We showed that our strategies are able to improve inference-time by 63% at the cost of performance drop of mere 3% for our problem-statement defined in evaluation section. These strategies brings down Vision Transformers detectors inference-time even less than traditional single-image based CNN detectors like FCOS. We recommend practitioners use these techniques to deploy Transformers based hefty multi-view networks on a budge-constrained robotic platform.
Pest counting, which predicts the number of pests in the early stage, is very important because it enables rapid pest control, reduces damage to crops, and improves productivity. In recent years, light traps have been increasingly used to lure and photograph pests for pest counting. However, pest images have a wide range of variability in pest appearance owing to severe occlusion, wide pose variation, and even scale variation. This makes pest counting more challenging. To address these issues, this study proposes a new pest counting model referred to as multiscale and deformable attention CenterNet (Mada-CenterNet) for internal low-resolution (LR) and high-resolution (HR) joint feature learning. Compared with the conventional CenterNet, the proposed Mada-CenterNet adopts a multiscale heatmap generation approach in a two-step fashion to predict LR and HR heatmaps adaptively learned to scale variations, that is, changes in the number of pests. In addition, to overcome the pose and occlusion problems, a new between-hourglass skip connection based on deformable and multiscale attention is designed to ensure internal LR and HR joint feature learning and incorporate geometric deformation, thereby resulting in an improved pest counting accuracy. Through experiments, the proposed Mada-CenterNet is verified to generate the HR heatmap more accurately and improve pest counting accuracy owing to multiscale heatmap generation, joint internal feature learning, and deformable and multiscale attention. In addition, the proposed model is confirmed to be effective in overcoming severe occlusions and variations in pose and scale. The experimental results show that the proposed model outperforms state-of-the-art crowd counting and object detection models.
Compared to general document analysis tasks, form document structure understanding and retrieval are challenging. Form documents are typically made by two types of authors; A form designer, who develops the form structure and keys, and a form user, who fills out form values based on the provided keys. Hence, the form values may not be aligned with the form designer's intention (structure and keys) if a form user gets confused. In this paper, we introduce Form-NLU, the first novel dataset for form structure understanding and its key and value information extraction, interpreting the form designer's intent and the alignment of user-written value on it. It consists of 857 form images, 6k form keys and values, and 4k table keys and values. Our dataset also includes three form types: digital, printed, and handwritten, which cover diverse form appearances and layouts. We propose a robust positional and logical relation-based form key-value information extraction framework. Using this dataset, Form-NLU, we first examine strong object detection models for the form layout understanding, then evaluate the key information extraction task on the dataset, providing fine-grained results for different types of forms and keys. Furthermore, we examine it with the off-the-shelf pdf layout extraction tool and prove its feasibility in real-world cases.
Large scale vision and language models can achieve impressive zero-shot recognition performance by mapping class specific text queries to image content. Two distinct challenges that remain however, are high sensitivity to the choice of handcrafted class names that define queries, and the difficulty of adaptation to new, smaller datasets. Towards addressing these problems, we propose to leverage available data to learn, for each class, an optimal word embedding as a function of the visual content. By learning new word embeddings on an otherwise frozen model, we are able to retain zero-shot capabilities for new classes, easily adapt models to new datasets, and adjust potentially erroneous, non-descriptive or ambiguous class names. We show that our solution can easily be integrated in image classification and object detection pipelines, yields significant performance gains in multiple scenarios and provides insights into model biases and labelling errors.
Monitoring the movement and actions of humans in video in real-time is an important task. We present a deep learning based algorithm for human action recognition for both RGB and thermal cameras. It is able to detect and track humans and recognize four basic actions (standing, walking, running, lying) in real-time on a notebook with a NVIDIA GPU. For this, it combines state of the art components for object detection (Scaled YoloV4), optical flow (RAFT) and pose estimation (EvoSkeleton). Qualitative experiments on a set of tunnel videos show that the proposed algorithm works robustly for both RGB and thermal video.
Bird's-Eye View (BEV) features are popular intermediate scene representations shared by the 3D backbone and the detector head in LiDAR-based object detectors. However, little research has been done to investigate how to incorporate additional supervision on the BEV features to improve proposal generation in the detector head, while still balancing the number of powerful 3D layers and efficient 2D network operations. This paper proposes a novel scene representation that encodes both the semantics and geometry of the 3D environment in 2D, which serves as a dense supervision signal for better BEV feature learning. The key idea is to use auxiliary networks to predict a combination of explicit and implicit semantic probabilities by exploiting their complementary properties. Extensive experiments show that our simple yet effective design can be easily integrated into most state-of-the-art 3D object detectors and consistently improves upon baseline models.
State-of-the-art 3D object detectors are usually trained on large-scale datasets with high-quality 3D annotations. However, such 3D annotations are often expensive and time-consuming, which may not be practical for real applications. A natural remedy is to adopt semi-supervised learning (SSL) by leveraging a limited amount of labeled samples and abundant unlabeled samples. Current pseudolabeling-based SSL object detection methods mainly adopt a teacher-student framework, with a single fixed threshold strategy to generate supervision signals, which inevitably brings confused supervision when guiding the student network training. Besides, the data augmentation of the point cloud in the typical teacher-student framework is too weak, and only contains basic down sampling and flip-and-shift (i.e., rotate and scaling), which hinders the effective learning of feature information. Hence, we address these issues by introducing a novel approach of Hierarchical Supervision and Shuffle Data Augmentation (HSSDA), which is a simple yet effective teacher-student framework. The teacher network generates more reasonable supervision for the student network by designing a dynamic dual-threshold strategy. Besides, the shuffle data augmentation strategy is designed to strengthen the feature representation ability of the student network. Extensive experiments show that HSSDA consistently outperforms the recent state-of-the-art methods on different datasets. The code will be released at https://github.com/azhuantou/HSSDA.
The main challenge of monocular 3D object detection is the accurate localization of 3D center. Motivated by a new and strong observation that this challenge can be remedied by a 3D-space local-grid search scheme in an ideal case, we propose a stage-wise approach, which combines the information flow from 2D-to-3D (3D bounding box proposal generation with a single 2D image) and 3D-to-2D (proposal verification by denoising with 3D-to-2D contexts) in a top-down manner. Specifically, we first obtain initial proposals from off-the-shelf backbone monocular 3D detectors. Then, we generate a 3D anchor space by local-grid sampling from the initial proposals. Finally, we perform 3D bounding box denoising at the 3D-to-2D proposal verification stage. To effectively learn discriminative features for denoising highly overlapped proposals, this paper presents a method of using the Perceiver I/O model to fuse the 3D-to-2D geometric information and the 2D appearance information. With the encoded latent representation of a proposal, the verification head is implemented with a self-attention module. Our method, named as MonoXiver, is generic and can be easily adapted to any backbone monocular 3D detectors. Experimental results on the well-established KITTI dataset and the challenging large-scale Waymo dataset show that MonoXiver consistently achieves improvement with limited computation overhead.
Domain shift significantly influences the performance of deep learning algorithms, particularly for object detection within volumetric 3D images. Annotated training data is essential for deep learning-based object detection. However, annotating densely packed objects is time-consuming and costly. Instead, we suggest training models on individually scanned objects, causing a domain shift between training and detection data. To address this challenge, we introduce the BugNIST dataset, comprising 9154 micro-CT volumes of 12 bug types and 388 volumes of tightly packed bug mixtures. This dataset is characterized by having objects with the same appearance in the source and target domains, which is uncommon for other benchmark datasets for domain shift. During training, individual bug volumes labeled by class are utilized, while testing employs mixtures with center point annotations and bug type labels. Together with the dataset, we provide a baseline detection analysis, with the aim of advancing the field of 3D object detection methods.
Since the introduction of deep learning, a wide scope of representation properties, such as decorrelation, whitening, disentanglement, rank, isotropy, and mutual information, have been studied to improve the quality of representation. However, manipulating such properties can be challenging in terms of implementational effectiveness and general applicability. To address these limitations, we propose to regularize von Neumann entropy~(VNE) of representation. First, we demonstrate that the mathematical formulation of VNE is superior in effectively manipulating the eigenvalues of the representation autocorrelation matrix. Then, we demonstrate that it is widely applicable in improving state-of-the-art algorithms or popular benchmark algorithms by investigating domain-generalization, meta-learning, self-supervised learning, and generative models. In addition, we formally establish theoretical connections with rank, disentanglement, and isotropy of representation. Finally, we provide discussions on the dimension control of VNE and the relationship with Shannon entropy. Code is available at: https://github.com/jaeill/CVPR23-VNE.
Although end-to-end video text spotting methods based on Transformer can model long-range dependencies and simplify the train process, it will lead to large computation cost with the increase of the frame size in the input video. Therefore, considering the resolution of ICDAR 2023 DSText is 1080 * 1920 and slicing the video frame into several areas will destroy the spatial correlation of text, we divided the small and dense text spotting into two tasks, text detection and tracking. For text detection, we adopt the PP-YOLOE-R which is proven effective in small object detection as our detection model. For text detection, we use the sort algorithm for high inference speed. Experiments on DSText dataset demonstrate that our method is competitive on small and dense text spotting.
The goal of open-vocabulary detection is to identify novel objects based on arbitrary textual descriptions. In this paper, we address open-vocabulary 3D point-cloud detection by a dividing-and-conquering strategy, which involves: 1) developing a point-cloud detector that can learn a general representation for localizing various objects, and 2) connecting textual and point-cloud representations to enable the detector to classify novel object categories based on text prompting. Specifically, we resort to rich image pre-trained models, by which the point-cloud detector learns localizing objects under the supervision of predicted 2D bounding boxes from 2D pre-trained detectors. Moreover, we propose a novel de-biased triplet cross-modal contrastive learning to connect the modalities of image, point-cloud and text, thereby enabling the point-cloud detector to benefit from vision-language pre-trained models,i.e.,CLIP. The novel use of image and vision-language pre-trained models for point-cloud detectors allows for open-vocabulary 3D object detection without the need for 3D annotations. Experiments demonstrate that the proposed method improves at least 3.03 points and 7.47 points over a wide range of baselines on the ScanNet and SUN RGB-D datasets, respectively. Furthermore, we provide a comprehensive analysis to explain why our approach works.
As the population grows and more land is being used for urbanization, ecosystems are disrupted by our roads and cars. This expansion of infrastructure cuts through wildlife territories, leading to many instances of Wildlife-Vehicle Collision (WVC). These instances of WVC are a global issue that is having a global socio-economic impact, resulting in billions of dollars in property damage and, at times, fatalities for vehicle occupants. In Saudi Arabia, this issue is similar, with instances of Camel-Vehicle Collision (CVC) being particularly deadly due to the large size of camels, which results in a 25% fatality rate [1]. The focus of this work is to test different object detection models on the task of detecting camels on the road. The Deep Learning (DL) object detection models used in the experiments are: Center Net, Efficient Det, Faster R-CNN, SSD, and YOLOv8. Results of the experiments show that YOLOv8 performed the best in terms of accuracy and was the most efficient in training. In the future, the plan is to expand on this work by developing a system to make countryside roads safer.
Multi-class cell detection and counting is an essential task for many pathological diagnoses. Manual counting is tedious and often leads to inter-observer variations among pathologists. While there exist multiple, general-purpose, deep learning-based object detection and counting methods, they may not readily transfer to detecting and counting cells in medical images, due to the limited data, presence of tiny overlapping objects, multiple cell types, severe class-imbalance, minute differences in size/shape of cells, etc. In response, we propose guided posterior regularization (DeGPR), which assists an object detector by guiding it to exploit discriminative features among cells. The features may be pathologist-provided or inferred directly from visual data. We validate our model on two publicly available datasets (CoNSeP and MoNuSAC), and on MuCeD, a novel dataset that we contribute. MuCeD consists of 55 biopsy images of the human duodenum for predicting celiac disease. We perform extensive experimentation with three object detection baselines on three datasets to show that DeGPR is model-agnostic, and consistently improves baselines obtaining up to 9% (absolute) mAP gains.
With advances in image recognition technology based on deep learning, automatic video analysis by Artificial Intelligence is becoming more widespread. As the amount of video used for image recognition increases, efficient compression methods for such video data are necessary. In general, when the image quality deteriorates due to image encoding, the image recognition accuracy also falls. Therefore, in this paper, we propose a neural-network-based approach to improve image recognition accuracy, especially the object detection accuracy by applying post-processing to the encoded video. Versatile Video Coding (VVC) will be used for the video compression method, since it is the latest video coding method with the best encoding performance. The neural network is trained using the features of YOLO-v7, the latest object detection model. By using VVC as the video coding method and YOLO-v7 as the detection model, high object detection accuracy is achieved even at low bit rates. Experimental results show that the combination of the proposed method and VVC achieves better coding performance than regular VVC in object detection accuracy.