In the past few years, mobile deep-learning deployment progressed by leaps and bounds, but solutions still struggle to accommodate its severe and fluctuating operational restrictions, which include bandwidth, latency, computation, and energy. In this work, we help to bridge that gap, introducing the first configurable solution for object detection that manages the triple communication-computation-accuracy trade-off with a single set of weights. Our solution shows state-of-the-art results on COCO-2017, adding only a minor penalty on the base EfficientDet-D2 architecture. Our design is robust to the choice of base architecture and compressor and should adapt well for future architectures.
There is a lot of ongoing research effort into developing different techniques for neural networks compression. However, the community lacks standardised evaluation metrics, which are key to identifying the most suitable compression technique for different applications. This paper reviews existing neural network compression evaluation metrics and implements them into a standardisation framework called NetZIP. We introduce two novel metrics to cover existing gaps of evaluation in the literature: 1) Compression and Hardware Agnostic Theoretical Speed (CHATS) and 2) Overall Compression Success (OCS). We demonstrate the use of NetZIP using two case studies on two different hardware platforms (a PC and a Raspberry Pi 4) focusing on object classification and object detection.
Recent works have shown that neural networks are vulnerable to carefully crafted adversarial examples (AE). By adding small perturbations to input images, AEs are able to make the victim model predicts incorrect outputs. Several research work in adversarial machine learning started to focus on the detection of AEs in autonomous driving. However, the existing studies either use preliminary assumption on outputs of detections or ignore the tracking system in the perception pipeline. In this paper, we firstly propose a novel distance metric for practical autonomous driving object detection outputs. Then, we bridge the gap between the current AE detection research and the real-world autonomous systems by providing a temporal detection algorithm, which takes the impact of tracking system into consideration. We perform evaluation on Berkeley Deep Drive (BDD) and CityScapes datasets to show how our approach outperforms existing single-frame-mAP based AE detections by increasing 17.76% accuracy of performance.
Soccer analytics is attracting increasing interest in academia and industry, thanks to the availability of data that describe all the spatio-temporal events that occur in each match. These events (e.g., passes, shots, fouls) are collected by human operators manually, constituting a considerable cost for data providers in terms of time and economic resources. In this paper, we describe PassNet, a method to recognize the most frequent events in soccer, i.e., passes, from video streams. Our model combines a set of artificial neural networks that perform feature extraction from video streams, object detection to identify the positions of the ball and the players, and classification of frame sequences as passes or not passes. We test PassNet on different scenarios, depending on the similarity of conditions to the match used for training. Our results show good classification results and significant improvement in the accuracy of pass detection with respect to baseline classifiers, even when the match's video conditions of the test and training sets are considerably different. PassNet is the first step towards an automated event annotation system that may break the time and the costs for event annotation, enabling data collections for minor and non-professional divisions, youth leagues and, in general, competitions whose matches are not currently annotated by data providers.
This study addresses the evolving challenges in urban traffic monitoring detection systems based on fisheye lens cameras by proposing a framework that improves the efficacy and accuracy of these systems. In the context of urban infrastructure and transportation management, advanced traffic monitoring systems have become critical for managing the complexities of urbanization and increasing vehicle density. Traditional monitoring methods, which rely on static cameras with narrow fields of view, are ineffective in dynamic urban environments, necessitating the installation of multiple cameras, which raises costs. Fisheye lenses, which were recently introduced, provide wide and omnidirectional coverage in a single frame, making them a transformative solution. However, issues such as distorted views and blurriness arise, preventing accurate object detection on these images. Motivated by these challenges, this study proposes a novel approach that combines a ransformer-based image enhancement framework and ensemble learning technique to address these challenges and improve traffic monitoring accuracy, making significant contributions to the future of intelligent traffic management systems. Our proposed methodological framework won 5th place in the 2024 AI City Challenge, Track 4, with an F1 score of 0.5965 on experimental validation data. The experimental results demonstrate the effectiveness, efficiency, and robustness of the proposed system. Our code is publicly available at https://github.com/daitranskku/AIC2024-TRACK4-TEAM15.
Infrared-visible object detection (IVOD) seeks to harness the complementary information in infrared and visible images, thereby enhancing the performance of detectors in complex environments. However, existing methods often neglect the frequency characteristics of complementary information, such as the abundant high-frequency details in visible images and the valuable low-frequency thermal information in infrared images, thus constraining detection performance. To solve this problem, we introduce a novel Frequency-Driven Feature Decomposition Network for IVOD, called FD2-Net, which effectively captures the unique frequency representations of complementary information across multimodal visual spaces. Specifically, we propose a feature decomposition encoder, wherein the high-frequency unit (HFU) utilizes discrete cosine transform to capture representative high-frequency features, while the low-frequency unit (LFU) employs dynamic receptive fields to model the multi-scale context of diverse objects. Next, we adopt a parameter-free complementary strengths strategy to enhance multimodal features through seamless inter-frequency recoupling. Furthermore, we innovatively design a multimodal reconstruction mechanism that recovers image details lost during feature extraction, further leveraging the complementary information from infrared and visible images to enhance overall representational capacity. Extensive experiments demonstrate that FD2-Net outperforms state-of-the-art (SOTA) models across various IVOD benchmarks, i.e. LLVIP (96.2% mAP), FLIR (82.9% mAP), and M3FD (83.5% mAP).
Interactive devices with touch screen have become commonly used in various aspects of daily life, which raises the demand for high production quality of touch screen glass. While it is desirable to develop effective defect detection technologies to optimize the automatic touch screen production lines, the development of these technologies suffers from the lack of publicly available datasets. To address this issue, we in this paper propose a dedicated touch screen glass defect dataset which includes seven types of defects and consists of 2504 images captured in various scenarios.All data are captured with professional acquisition equipment on the fixed workstation. Additionally, we benchmark the CNN- and Transformer-based object detection frameworks on the proposed dataset to demonstrate the challenges of defect detection on high-resolution images. Dataset and related code will be available at https://github.com/Yangr116/SSGDataset.
This paper presents a comprehensive solution to address the critical challenge of liquid leaks in the oil and gas industry, leveraging advanced computer vision and deep learning methodologies. Employing You Only Look Once (YOLO) and Real-Time Detection Transformer (RT DETR) models, our project focuses on enhancing early identification of liquid leaks in key infrastructure components such as pipelines, pumps, and tanks. Through the integration of surveillance thermal cameras and sensors, the combined YOLO and RT DETR models demonstrate remarkable efficacy in the continuous monitoring and analysis of visual data within oil and gas facilities. YOLO's real-time object detection capabilities swiftly recognize leaks and their patterns, while RT DETR excels in discerning specific leak-related features, particularly in thermal images. This approach significantly improves the accuracy and speed of leak detection, ultimately mitigating environmental and financial risks associated with liquid leaks.
T1-T2 상관 MRI는 생물학적 모델을 가정하지 않고도 미세구조 구획을 정량화할 수 있는 큰 잠재력을 지닌다. 전통적으로 복셀별 전체 T1-T2 스펙트럼은 조건이 나쁘고 시간이 많이 소요되는 역라플라스 변환(Inverse Laplace Transform)으로 복원된다. 본 연구는 이 문제를 객체 검출 과제로 바라볼 것을 제안한다. 검출 트랜스포머(Detection Transformer, DETR)가 64개의 TI/TE 신호를 입력받아 T1, T2, 신호 분율로 지정되는 구획 집합을 직접 예측한다. 시뮬레이션 데이터에서 평균 정밀도 mAP@10 76.8%를 달성하여, 기존 모델 설계(53.8%)와 정규화된 NNLS(66.1%)를 모두 능가한다. 또한 DETR 방식이 구획 수, 구획별 신호 분율, SNR에 걸친 거의 모든 설정에서 NNLS보다 우수함을 보인다. 복원 속도는 NNLS보다 수 order of magnitude 빠르다. 결과는 공개 코드(github.com/SebastianEndtTHI/t1t2-detr)로 완전히 재현할 수 있다.
Visual SLAM systems targeting static scenes have been developed with satisfactory accuracy and robustness. Dynamic 3D object tracking has then become a significant capability in visual SLAM with the requirement of understanding dynamic surroundings in various scenarios including autonomous driving, augmented and virtual reality. However, performing dynamic SLAM solely with monocular images remains a challenging problem due to the difficulty of associating dynamic features and estimating their positions. In this paper, we present MOTSLAM, a dynamic visual SLAM system with the monocular configuration that tracks both poses and bounding boxes of dynamic objects. MOTSLAM first performs multiple object tracking (MOT) with associated both 2D and 3D bounding box detection to create initial 3D objects. Then, neural-network-based monocular depth estimation is applied to fetch the depth of dynamic features. Finally, camera poses, object poses, and both static, as well as dynamic map points, are jointly optimized using a novel bundle adjustment. Our experiments on the KITTI dataset demonstrate that our system has reached best performance on both camera ego-motion and object tracking on monocular dynamic SLAM.
In this paper, we propose a novel deep learning based approach for identifying co-occurring objects in conjunction with base objects in multilabel object categories. Nowadays, with the advancement in computer vision based techniques we need to know about co-occurring objects with respect to base object for various purposes. The pipeline of the proposed work is composed of two stages: in the first stage of the proposed model we detect all the bounding boxes present in the image and their corresponding labels, then in the second stage we perform co-occurrence matrix analysis. In co-occurrence matrix analysis, we set base classes based on the maximum occurrences of the labels and build association rules and generate frequent patterns. These frequent patterns will show base classes and their corresponding co-occurring classes. We performed our experiments on two publicly available datasets: Pascal VOC and MS-COCO. The experimental results on public benchmark dataset is reported in Sec 4. Further we extend this work by considering all frequently objects as unlabeled and what if they are occluded as well.
Conventional radar feature extraction faces limitations due to low spatial resolution, noise, multipath reflection, the presence of ghost targets, and motion blur. Such limitations can be exacerbated by nonlinear object motion, particularly from an ego-centric viewpoint. It becomes evident that to address these challenges, the key lies in exploiting temporal feature relation over an extended horizon and enforcing spatial motion consistency for effective association. To this end, this paper proposes SIRA (Scalable Inter-frame Relation and Association) with two designs. First, inspired by Swin Transformer, we introduce extended temporal relation, generalizing the existing temporal relation layer from two consecutive frames to multiple inter-frames with temporally regrouped window attention for scalability. Second, we propose motion consistency track with the concept of a pseudo-tracklet generated from observational data for better trajectory prediction and subsequent object association. Our approach achieves 58.11 [email omitted] for oriented object detection and 47.79 MOTA for multiple object tracking on the Radiate dataset, surpassing previous state-of-the-art by a margin of +4.11 [email omitted] and +9.94 MOTA, respectively.
Since detecting and recognizing individual human or object are not adequate to understand the visual world, learning how humans interact with surrounding objects becomes a core technology. However, convolution operations are weak in depicting visual interactions between the instances since they only build blocks that process one local neighborhood at a time. To address this problem, we learn from human perception in observing HOIs to introduce a two-stage trainable reasoning mechanism, referred to as GID block. GID block breaks through the local neighborhoods and captures long-range dependency of pixels both in global-level and instance-level from the scene to help detecting interactions between instances. Furthermore, we conduct a multi-stream network called GID-Net, which is a human-object interaction detection framework consisting of a human branch, an object branch and an interaction branch. Semantic information in global-level and local-level are efficiently reasoned and aggregated in each of the branches. We have compared our proposed GID-Net with existing state-of-the-art methods on two public benchmarks, including V-COCO and HICO-DET. The results have showed that GID-Net outperforms the existing best-performing methods on both the above two benchmarks, validating its efficacy in detecting human-object interactions.
3D dense captioning is a task to localize objects in a 3D scene and generate descriptive sentences for each object. Recent approaches in 3D dense captioning have adopted transformer encoder-decoder frameworks from object detection to build an end-to-end pipeline without hand-crafted components. However, these approaches struggle with contradicting objectives where a single query attention has to simultaneously view both the tightly localized object regions and contextual environment. To overcome this challenge, we introduce SIA (See-It-All), a transformer pipeline that engages in 3D dense captioning with a novel paradigm called late aggregation. SIA simultaneously decodes two sets of queries-context query and instance query. The instance query focuses on localization and object attribute descriptions, while the context query versatilely captures the region-of-interest of relationships between multiple objects or with the global scene, then aggregated afterwards (i.e., late aggregation) via simple distance-based measures. To further enhance the quality of contextualized caption generation, we design a novel aggregator to generate a fully informed caption based on the surrounding context, the global environment, and object instances. Extensive experiments on two of the most widely-used 3D dense captioning datasets demonstrate that our proposed method achieves a significant improvement over prior methods.
We address the problem of action detection in videos. Driven by the latest progress in object detection from 2D images, we build action models using rich feature hierarchies derived from shape and kinematic cues. We incorporate appearance and motion in two ways. First, starting from image region proposals we select those that are motion salient and thus are more likely to contain the action. This leads to a significant reduction in the number of regions being processed and allows for faster computations. Second, we extract spatio-temporal feature representations to build strong classifiers using Convolutional Neural Networks. We link our predictions to produce detections consistent in time, which we call action tubes. We show that our approach outperforms other techniques in the task of action detection.
This study developed a traffic sign detection and recognition algorithm based on the RetinaNet. Two main aspects were revised to improve the detection of traffic signs: image cropping to address the issue of large image and small traffic signs; and using more anchors with various scales to detect traffic signs with different sizes and shapes. The proposed algorithm was trained and tested in a series of autonomous driving front-view images in a virtual simulation environment. Results show that the algorithm performed extremely well under good illumination and weather conditions. Its drawbacks are that it sometimes failed to detect object under bad weather conditions like snow and failed to distinguish speed limits signs with different limit values.
In this study, we present an analysis of model-based ensemble learning for 3D point-cloud object classification and detection. An ensemble of multiple model instances is known to outperform a single model instance, but there is little study of the topic of ensemble learning for 3D point clouds. First, an ensemble of multiple model instances trained on the same part of the $\textit{ModelNet40}$ dataset was tested for seven deep learning, point cloud-based classification algorithms: $\textit{PointNet}$, $\textit{PointNet++}$, $\textit{SO-Net}$, $\textit{KCNet}$, $\textit{DeepSets}$, $\textit{DGCNN}$, and $\textit{PointCNN}$. Second, the ensemble of different architectures was tested. Results of our experiments show that the tested ensemble learning methods improve over state-of-the-art on the $\textit{ModelNet40}$ dataset, from $92.65\%$ to $93.64\%$ for the ensemble of single architecture instances, $94.03\%$ for two different architectures, and $94.15\%$ for five different architectures. We show that the ensemble of two models with different architectures can be as effective as the ensemble of 10 models with the same architecture. Third, a study on classic bagging i.e. with different subsets used for training multiple model instances) was tested and sources of ensemble accuracy growth were investigated for best-performing architecture, i.e. $\textit{SO-Net}$. We also investigate the ensemble learning of $\textit{Frustum PointNet}$ approach in the task of 3D object detection, increasing the average precision of 3D box detection on the $\textit{KITTI}$ dataset from $63.1\%$ to $66.5\%$ using only three model instances. We measure the inference time of all 3D classification architectures on a $\textit{Nvidia Jetson TX2}$, a common embedded computer for mobile robots, to allude to the use of these models in real-life applications.
Object detection has achieved substantial progress in the last decade. However, detecting novel classes with only few samples remains challenging, since deep learning under low data regime usually leads to a degraded feature space. Existing works employ a holistic fine-tuning paradigm to tackle this problem, where the model is first pre-trained on all base classes with abundant samples, and then it is used to carve the novel class feature space. Nonetheless, this paradigm is still imperfect. Durning fine-tuning, a novel class may implicitly leverage the knowledge of multiple base classes to construct its feature space, which induces a scattered feature space, hence violating the inter-class separability. To overcome these obstacles, we propose a two-step fine-tuning framework, Few-shot object detection via Association and DIscrimination (FADI), which builds up a discriminative feature space for each novel class with two integral steps. 1) In the association step, in contrast to implicitly leveraging multiple base classes, we construct a compact novel class feature space via explicitly imitating a specific base class feature space. Specifically, we associate each novel class with a base class according to their semantic similarity. After that, the feature space of a novel class can readily imitate the well-trained feature space of the associated base class. 2) In the discrimination step, to ensure the separability between the novel classes and associated base classes, we disentangle the classification branches for base and novel classes. To further enlarge the inter-class separability between all classes, a set-specialized margin loss is imposed. Extensive experiments on Pascal VOC and MS-COCO datasets demonstrate FADI achieves new SOTA performance, significantly improving the baseline in any shot/split by +18.7. Notably, the advantage is most announced on extremely few-shot scenarios.
Modern object detectors are vulnerable to adversarial examples, which may bring risks to real-world applications. The sparse attack is an important task which, compared with the popular adversarial perturbation on the whole image, needs to select the potential pixels that is generally regularized by an $\ell_0$-norm constraint, and simultaneously optimize the corresponding texture. The non-differentiability of $\ell_0$ norm brings challenges and many works on attacking object detection adopted manually-designed patterns to address them, which are meaningless and independent of objects, and therefore lead to relatively poor attack performance. In this paper, we propose Adversarial Semantic Contour (ASC), an MAP estimate of a Bayesian formulation of sparse attack with a deceived prior of object contour. The object contour prior effectively reduces the search space of pixel selection and improves the attack by introducing more semantic bias. Extensive experiments demonstrate that ASC can corrupt the prediction of 9 modern detectors with different architectures (\e.g., one-stage, two-stage and Transformer) by modifying fewer than 5\% of the pixels of the object area in COCO in white-box scenario and around 10\% of those in black-box scenario. We further extend the attack to datasets for autonomous driving systems to verify the effectiveness. We conclude with cautions about contour being the common weakness of object detectors with various architecture and the care needed in applying them in safety-sensitive scenarios.
We present Laneformer, a conceptually simple yet powerful transformer-based architecture tailored for lane detection that is a long-standing research topic for visual perception in autonomous driving. The dominant paradigms rely on purely CNN-based architectures which often fail in incorporating relations of long-range lane points and global contexts induced by surrounding objects (e.g., pedestrians, vehicles). Inspired by recent advances of the transformer encoder-decoder architecture in various vision tasks, we move forwards to design a new end-to-end Laneformer architecture that revolutionizes the conventional transformers into better capturing the shape and semantic characteristics of lanes, with minimal overhead in latency. First, coupling with deformable pixel-wise self-attention in the encoder, Laneformer presents two new row and column self-attention operations to efficiently mine point context along with the lane shapes. Second, motivated by the appearing objects would affect the decision of predicting lane segments, Laneformer further includes the detected object instances as extra inputs of multi-head attention blocks in the encoder and decoder to facilitate the lane point detection by sensing semantic contexts. Specifically, the bounding box locations of objects are added into Key module to provide interaction with each pixel and query while the ROI-aligned features are inserted into Value module. Extensive experiments demonstrate our Laneformer achieves state-of-the-art performances on CULane benchmark, in terms of 77.1% F1 score. We hope our simple and effective Laneformer will serve as a strong baseline for future research in self-attention models for lane detection.