Knowledge Distillation (KD) is a well-known training paradigm in deep neural networks where knowledge acquired by a large teacher model is transferred to a small student. KD has proven to be an effective technique to significantly improve the student's performance for various tasks including object detection. As such, KD techniques mostly rely on guidance at the intermediate feature level, which is typically implemented by minimizing an lp-norm distance between teacher and student activations during training. In this paper, we propose a replacement for the pixel-wise independent lp-norm based on the structural similarity (SSIM). By taking into account additional contrast and structural cues, feature importance, correlation and spatial dependence in the feature space are considered in the loss formulation. Extensive experiments on MSCOCO demonstrate the effectiveness of our method across different training schemes and architectures. Our method adds only little computational overhead, is straightforward to implement and at the same time it significantly outperforms the standard lp-norms. Moreover, more complex state-of-the-art KD methods using attention-based sampling mechanisms are outperformed, including a +3.5 AP gain using a Faster R-CNN R-50 compared to a vanilla model.
As AI algorithms increasingly participate in daily activities, it becomes critical to ascertain whether the agents we interact with are human or not. To address this question, we turn to the Turing test and systematically benchmark current AIs in their abilities to imitate humans in three language tasks (Image captioning, Word association, and Conversation) and three vision tasks (Object detection, Color estimation, and Attention prediction). The experiments involved 549 human agents plus 26 AI agents for dataset creation, and 1,126 human judges plus 10 AI judges, in 25,650 Turing-like tests. The results reveal that current AIs are not far from being able to impersonate humans in complex language and vision challenges. While human judges were often deceived, simple AI judges outperformed human judges in distinguishing human answers from AI answers. The results of imitation tests are only minimally correlated with standard performance metrics in AI. Thus, evaluating whether a machine can pass as a human constitutes an important independent test to evaluate AI algorithms. The curated, large-scale, Turing datasets introduced here and their evaluation metrics provide new benchmarks and insights to assess whether an agent is human or not and emphasize the relevance of rigorous, systematic, and quantitative imitation tests in these and other AI domains.
We propose integrally pre-trained transformer pyramid network (iTPN), towards jointly optimizing the network backbone and the neck, so that transfer gap between representation models and downstream tasks is minimal. iTPN is born with two elaborated designs: 1) The first pre-trained feature pyramid upon vision transformer (ViT). 2) Multi-stage supervision to the feature pyramid using masked feature modeling (MFM). iTPN is updated to Fast-iTPN, reducing computational memory overhead and accelerating inference through two flexible designs. 1) Token migration: dropping redundant tokens of the backbone while replenishing them in the feature pyramid without attention operations. 2) Token gathering: reducing computation cost caused by global attention by introducing few gathering tokens. The base/large-level Fast-iTPN achieve 88.75%/89.5% top-1 accuracy on ImageNet-1K. With 1x training schedule using DINO, the base/large-level Fast-iTPN achieves 58.4%/58.8% box AP on COCO object detection, and a 57.5%/58.7% mIoU on ADE20K semantic segmentation using MaskDINO. Fast-iTPN can accelerate the inference procedure by up to 70%, with negligible performance loss, demonstrating the potential to be a powerful backbone for downstream vision tasks. The code is available at: github.com/sunsmarterjie/iTPN.
Joint bleeding is a common condition for people with hemophilia and, if untreated, can result in hemophilic arthropathy. Ultrasound imaging has recently emerged as an effective tool to diagnose joint recess distension caused by joint bleeding. However, no computer-aided diagnosis tool exists to support the practitioner in the diagnosis process. This paper addresses the problem of automatically detecting the recess and assessing whether it is distended in knee ultrasound images collected in patients with hemophilia. After framing the problem, we propose two different approaches: the first one adopts a one-stage object detection algorithm, while the second one is a multi-task approach with a classification and a detection branch. The experimental evaluation, conducted with $483$ annotated images, shows that the solution based on object detection alone has a balanced accuracy score of $0.74$ with a mean IoU value of $0.66$, while the multi-task approach has a higher balanced accuracy value ($0.78$) at the cost of a slightly lower mean IoU value.
The camouflaged object detection (COD) task aims to identify and segment objects that blend into the background due to their similar color or texture. Despite the inherent difficulties of the task, COD has gained considerable attention in several fields, such as medicine, life-saving, and anti-military fields. In this paper, we propose a novel solution called the Deformable Point Sampling network (DPS-Net) to address the challenges associated with COD. The proposed DPS-Net utilizes a Deformable Point Sampling transformer (DPS transformer) that can effectively capture sparse local boundary information of significant object boundaries in COD using a deformable point sampling method. Moreover, the DPS transformer demonstrates robust COD performance by extracting contextual features for target object localization through integrating rough global positional information of objects with boundary local information. We evaluate our method on three prominent datasets and achieve state-of-the-art performance. Our results demonstrate the effectiveness of the proposed method through comparative experiments.
Implicit Neural Representations (INRs) are powerful to parameterize continuous signals in computer vision. However, almost all INRs methods are limited to low-level tasks, e.g., image/video compression, super-resolution, and image generation. The questions on how to explore INRs to high-level tasks and deep networks are still under-explored. Existing INRs methods suffer from two problems: 1) narrow theoretical definitions of INRs are inapplicable to high-level tasks; 2) lack of representation capabilities to deep networks. Motivated by the above facts, we reformulate the definitions of INRs from a novel perspective and propose an innovative Implicit Neural Representation Network (INRN), which is the first study of INRs to tackle both low-level and high-level tasks. Specifically, we present three key designs for basic blocks in INRN along with two different stacking ways and corresponding loss functions. Extensive experiments with analysis on both low-level tasks (image fitting) and high-level vision tasks (image classification, object detection, instance segmentation) demonstrate the effectiveness of the proposed method.
3D object detection received increasing attention in autonomous driving recently. Objects in 3D scenes are distributed with diverse orientations. Ordinary detectors do not explicitly model the variations of rotation and reflection transformations. Consequently, large networks and extensive data augmentation are required for robust detection. Recent equivariant networks explicitly model the transformation variations by applying shared networks on multiple transformed point clouds, showing great potential in object geometry modeling. However, it is difficult to apply such networks to 3D object detection in autonomous driving due to its large computation cost and slow reasoning speed. In this work, we present TED, an efficient Transformation-Equivariant 3D Detector to overcome the computation cost and speed issues. TED first applies a sparse convolution backbone to extract multi-channel transformation-equivariant voxel features; and then aligns and aggregates these equivariant features into lightweight and compact representations for high-performance 3D object detection. On the highly competitive KITTI 3D car detection leaderboard, TED ranked 1st among all submissions with competitive efficiency.
Semi-supervised Learning (SSL) has received increasing attention in autonomous driving to reduce the enormous burden of 3D annotation. In this paper, we propose UpCycling, a novel SSL framework for 3D object detection with zero additional raw-level point cloud: learning from unlabeled de-identified intermediate features (i.e., smashed data) to preserve privacy. Since these intermediate features are naturally produced by the inference pipeline, no additional computation is required on autonomous vehicles. However, generating effective consistency loss for unlabeled feature-level scene turns out to be a critical challenge. The latest SSL frameworks for 3D object detection that enforce consistency regularization between different augmentations of an unlabeled raw-point scene become detrimental when applied to intermediate features. To solve the problem, we introduce a novel combination of hybrid pseudo labels and feature-level Ground Truth sampling (F-GT), which safely augments unlabeled multi-type 3D scene features and provides high-quality supervision. We implement UpCycling on two representative 3D object detection models: SECOND-IoU and PV-RCNN. Experiments on widely-used datasets (Waymo, KITTI, and Lyft) verify that UpCycling outperforms other augmentation methods applied at the feature level. In addition, while preserving privacy, UpCycling performs better or comparably to the state-of-the-art methods that utilize raw-level unlabeled data in both domain adaptation and partial-label scenarios.
This paper does not attempt to design a state-of-the-art method for visual recognition but investigates a more efficient way to make use of convolutions to encode spatial features. By comparing the design principles of the recent convolutional neural networks ConvNets) and Vision Transformers, we propose to simplify the self-attention by leveraging a convolutional modulation operation. We show that such a simple approach can better take advantage of the large kernels (>=7x7) nested in convolutional layers. We build a family of hierarchical ConvNets using the proposed convolutional modulation, termed Conv2Former. Our network is simple and easy to follow. Experiments show that our Conv2Former outperforms existent popular ConvNets and vision Transformers, like Swin Transformer and ConvNeXt in all ImageNet classification, COCO object detection and ADE20k semantic segmentation.
Recent LSS-based multi-view 3D object detection has made tremendous progress, by processing the features in Brid-Eye-View (BEV) via the convolutional detector. However, the typical convolution ignores the radial symmetry of the BEV features and increases the difficulty of the detector optimization. To preserve the inherent property of the BEV features and ease the optimization, we propose an azimuth-equivariant convolution (AeConv) and an azimuth-equivariant anchor. The sampling grid of AeConv is always in the radial direction, thus it can learn azimuth-invariant BEV features. The proposed anchor enables the detection head to learn predicting azimuth-irrelevant targets. In addition, we introduce a camera-decoupled virtual depth to unify the depth prediction for the images with different camera intrinsic parameters. The resultant detector is dubbed Azimuth-equivariant Detector (AeDet). Extensive experiments are conducted on nuScenes, and AeDet achieves a 62.0% NDS, surpassing the recent multi-view 3D object detectors such as PETRv2 and BEVDepth by a large margin. Project page: https://fcjian.github.io/aedet.
State-of-the-art machine-learning methods for event cameras treat events as dense representations and process them with conventional deep neural networks. Thus, they fail to maintain the sparsity and asynchronous nature of event data, thereby imposing significant computation and latency constraints on downstream systems. A recent line of work tackles this issue by modeling events as spatiotemporally evolving graphs that can be efficiently and asynchronously processed using graph neural networks. These works showed impressive computation reductions, yet their accuracy is still limited by the small scale and shallow depth of their network, both of which are required to reduce computation. In this work, we break this glass ceiling by introducing several architecture choices which allow us to scale the depth and complexity of such models while maintaining low computation. On object detection tasks, our smallest model shows up to 3.7 times lower computation, while outperforming state-of-the-art asynchronous methods by 7.4 mAP. Even when scaling to larger model sizes, we are 13% more efficient than state-of-the-art while outperforming it by 11.5 mAP. As a result, our method runs 3.7 times faster than a dense graph neural network, taking only 8.4 ms per forward pass. This opens the door to efficient, and accurate object detection in edge-case scenarios.
Crowdedness caused by overlapping among similar objects is a ubiquitous challenge in the field of 2D visual object detection. In this paper, we first underline two main effects of the crowdedness issue: 1) IoU-confidence correlation disturbances (ICD) and 2) confused de-duplication (CDD). Then we explore a pathway of cracking these nuts from the perspective of data augmentation. Primarily, a particular copy-paste scheme is proposed towards making crowded scenes. Based on this operation, we first design a "consensus learning" method to further resist the ICD problem and then find out the pasting process naturally reveals a pseudo "depth" of object in the scene, which can be potentially used for alleviating CDD dilemma. Both methods are derived from magical using of the copy-pasting without extra cost for hand-labeling. Experiments show that our approach can easily improve the state-of-the-art detector in typical crowded detection task by more than 2% without any bells and whistles. Moreover, this work can outperform existing data augmentation strategies in crowded scenario.
There has been a significant increase in the adoption of technology in cricket recently. This trend has created the problem of duplicate work being done in similar computer vision-based research works. Our research tries to solve one of these problems by segmenting ball deliveries in a cricket broadcast using deep learning models, MobileNet and YOLO, thus enabling researchers to use our work as a dataset for their research. The output from our research can be used by cricket coaches and players to analyze ball deliveries which are played during the match. This paper presents an approach to segment and extract video shots in which only the ball is being delivered. The video shots are a series of continuous frames that make up the whole scene of the video. Object detection models are applied to reach a high level of accuracy in terms of correctly extracting video shots. The proof of concept for building large datasets of video shots for ball deliveries is proposed which paves the way for further processing on those shots for the extraction of semantics. Ball tracking in these video shots is also done using a separate RetinaNet model as a sample of the usefulness of the proposed dataset. The position on the cricket pitch where the ball lands is also extracted by tracking the ball along the y-axis. The video shot is then classified as a full-pitched, good-length or short-pitched delivery.
Images of spacecraft photographed from other spacecraft operating in outer space are difficult to come by, especially at a scale typically required for deep learning tasks. Semantic image segmentation, object detection and localization, and pose estimation are well researched areas with powerful results for many applications, and would be very useful in autonomous spacecraft operation and rendezvous. However, recent studies show that these strong results in broad and common domains may generalize poorly even to specific industrial applications on earth. To address this, we propose a method for generating synthetic image data that are labelled for semantic segmentation, generalizable to other tasks, and provide a prototype synthetic image dataset consisting of 2D monocular images of unmanned spacecraft, in order to enable further research in the area of autonomous spacecraft rendezvous. We also present a strong benchmark result (S{\o}rensen-Dice coefficient 0.8723) on these synthetic data, suggesting that it is feasible to train well-performing image segmentation models for this task, especially if the target spacecraft and its configuration are known.
In a cyber-physical system such as an autonomous vehicle (AV), machine learning (ML) models can be used to navigate and identify objects that may interfere with the vehicle's operation. However, ML models are unlikely to make accurate decisions when presented with data outside their training distribution. Out-of-distribution (OOD) detection can act as a safety monitor for ML models by identifying such samples at run time. However, in safety critical systems like AVs, OOD detection needs to satisfy real-time constraints in addition to functional requirements. In this demonstration, we use a mobile robot as a surrogate for an AV and use an OOD detector to identify potentially hazardous samples. The robot navigates a miniature town using image data and a YOLO object detection network. We show that our OOD detector is capable of identifying OOD images in real-time on an embedded platform concurrently performing object detection and lane following. We also show that it can be used to successfully stop the vehicle in the presence of unknown, novel samples.
Object detection is a computer vision task of predicting a set of bounding boxes and category labels for each object of interest in a given image. The category is related to a linguistic symbol such as 'dog' or 'person' and there should be relationships among them. However the object detector only learns to classify the categories and does not treat them as the linguistic symbols. Multi-modal models often use the pre-trained object detector to extract object features from the image, but the models are separated from the detector and the extracted visual features does not change with their linguistic input. We rethink the object detection as a vision-and-language reasoning task. We then propose targeted detection task, where detection targets are given by a natural language and the goal of the task is to detect only all the target objects in a given image. There are no detection if the target is not given. Commonly used modern object detectors have many hand-designed components like anchor and it is difficult to fuse the textual inputs into the complex pipeline. We thus propose Language-Targeted Detector (LTD) for the targeted detection based on a recently proposed Transformer-based detector. LTD is a encoder-decoder architecture and our conditional decoder allows the model to reason about the encoded image with the textual input as the linguistic context. We evaluate detection performances of LTD on COCO object detection dataset and also show that our model improves the detection results with the textual input grounding to the visual object.
Bird-eye-view (BEV) based methods have made great progress recently in multi-view 3D detection task. Comparing with BEV based methods, sparse based methods lag behind in performance, but still have lots of non-negligible merits. To push sparse 3D detection further, in this work, we introduce a novel method, named Sparse4D, which does the iterative refinement of anchor boxes via sparsely sampling and fusing spatial-temporal features. (1) Sparse 4D Sampling: for each 3D anchor, we assign multiple 4D keypoints, which are then projected to multi-view/scale/timestamp image features to sample corresponding features; (2) Hierarchy Feature Fusion: we hierarchically fuse sampled features of different view/scale, different timestamp and different keypoints to generate high-quality instance feature. In this way, Sparse4D can efficiently and effectively achieve 3D detection without relying on dense view transformation nor global attention, and is more friendly to edge devices deployment. Furthermore, we introduce an instance-level depth reweight module to alleviate the ill-posed issue in 3D-to-2D projection. In experiment, our method outperforms all sparse based methods and most BEV based methods on detection task in the nuScenes dataset.
Objection detection (OD) has been one of the most fundamental tasks in computer vision. Recent developments in deep learning have pushed the performance of image OD to new heights by learning-based, data-driven approaches. On the other hand, video OD remains less explored, mostly due to much more expensive data annotation needs. At the same time, multi-object tracking (MOT) which requires reasoning about track identities and spatio-temporal trajectories, shares similar spirits with video OD. However, most MOT datasets are class-specific (e.g., person-annotated only), which constrains a model's flexibility to perform tracking on other objects. We propose TrIVD (Tracking and Image-Video Detection), the first framework that unifies image OD, video OD, and MOT within one end-to-end model. To handle the discrepancies and semantic overlaps of category labels across datasets, TrIVD formulates detection/tracking as grounding and reasons about object categories via visual-text alignments. The unified formulation enables cross-dataset, multi-task training, and thus equips TrIVD with the ability to leverage frame-level features, video-level spatio-temporal relations, as well as track identity associations. With such joint training, we can now extend the knowledge from OD data, that comes with much richer object category annotations, to MOT and achieve zero-shot tracking capability. Experiments demonstrate that multi-task co-trained TrIVD outperforms single-task baselines across all image/video OD and MOT tasks. We further set the first baseline on the new task of zero-shot tracking.
For 3D object detection, labeling lidar point cloud is difficult, so data augmentation is an important module to make full use of precious annotated data. As a widely used data augmentation method, GT-sample effectively improves detection performance by inserting groundtruths into the lidar frame during training. However, these samples are often placed in unreasonable areas, which misleads model to learn the wrong context information between targets and backgrounds. To address this problem, in this paper, we propose a context-aware data augmentation method (CA-aug) , which ensures the reasonable placement of inserted objects by calculating the "Validspace" of the lidar point cloud. CA-aug is lightweight and compatible with other augmentation methods. Compared with the GT-sample and the similar method in Lidar-aug(SOTA), it brings higher accuracy to the existing detectors. We also present an in-depth study of augmentation methods for the range-view-based(RV-based) models and find that CA-aug can fully exploit the potential of RV-based networks. The experiment on KITTI val split shows that CA-aug can improve the mAP of the test model by 8%.
Automatic security inspection relying on computer vision technology is a challenging task in real-world scenarios due to many factors, such as intra-class variance, class imbalance, and occlusion. Most previous methods rarely touch the cases where the prohibited items are deliberately hidden in messy objects because of the scarcity of large-scale datasets, hindering their applications. To address this issue and facilitate related research, we present a large-scale dataset, named PIDray, which covers various cases in real-world scenarios for prohibited item detection, especially for deliberately hidden items. In specific, PIDray collects 124,486 X-ray images for $12$ categories of prohibited items, and each image is manually annotated with careful inspection, which makes it, to our best knowledge, to largest prohibited items detection dataset to date. Meanwhile, we propose a general divide-and-conquer pipeline to develop baseline algorithms on PIDray. Specifically, we adopt the tree-like structure to suppress the influence of the long-tailed issue in the PIDray dataset, where the first course-grained node is tasked with the binary classification to alleviate the influence of head category, while the subsequent fine-grained node is dedicated to the specific tasks of the tail categories. Based on this simple yet effective scheme, we offer strong task-specific baselines across object detection, instance segmentation, and multi-label classification tasks and verify the generalization ability on common datasets (e.g., COCO and PASCAL VOC). Extensive experiments on PIDray demonstrate that the proposed method performs favorably against current state-of-the-art methods, especially for deliberately hidden items. Our benchmark and codes will be released at https://github.com/lutao2021/PIDray.