The application of cross-dataset training in object detection tasks is complicated because the inconsistency in the category range across datasets transforms fully supervised learning into semi-supervised learning. To address this problem, recent studies focus on the generation of high-quality missing annotations. In this study, we first point out that it is not enough to generate high-quality annotations using a single model, which only looks once for annotations. Through detailed experimental analyses, we further conclude that hard-label training is conducive to generating high-recall annotations, while soft-label training tends to obtain high-precision annotations. Inspired by the aspects mentioned above, we propose a dynamic supervisor framework that updates the annotations multiple times through multiple-updated submodels trained using hard and soft labels. In the final generated annotations, both recall and precision improve significantly through the integration of hard-label training with soft-label training. Extensive experiments conducted on various dataset combination settings support our analyses and demonstrate the superior performance of the proposed dynamic supervisor.
This technical report presents the 1st place winning solution for the Waymo Open Dataset 3D semantic segmentation challenge 2022. Our network, termed LidarMultiNet, unifies the major LiDAR perception tasks such as 3D semantic segmentation, object detection, and panoptic segmentation in a single framework. At the core of LidarMultiNet is a strong 3D voxel-based encoder-decoder network with a novel Global Context Pooling (GCP) module extracting global contextual features from a LiDAR frame to complement its local features. An optional second stage is proposed to refine the first-stage segmentation or generate accurate panoptic segmentation results. Our solution achieves a mIoU of 71.13 and is the best for most of the 22 classes on the Waymo 3D semantic segmentation test set, outperforming all the other 3D semantic segmentation methods on the official leaderboard. We demonstrate for the first time that major LiDAR perception tasks can be unified in a single strong network that can be trained end-to-end.
Adversarial patches are images designed to fool otherwise well-performing neural network-based computer vision models. Although these attacks were initially conceived of and studied digitally, in that the raw pixel values of the image were perturbed, recent work has demonstrated that these attacks can successfully transfer to the physical world. This can be accomplished by printing out the patch and adding it into scenes of newly captured images or video footage. In this work we further test the efficacy of adversarial patch attacks in the physical world under more challenging conditions. We consider object detection models trained on overhead imagery acquired through aerial or satellite cameras, and we test physical adversarial patches inserted into scenes of a desert environment. Our main finding is that it is far more difficult to successfully implement the adversarial patch attacks under these conditions than in the previously considered conditions. This has important implications for AI safety as the real-world threat posed by adversarial examples may be overstated.
Camouflaged Object Detection (COD) is a challenging task in computer vision due to the high similarity between camouflaged objects and their surroundings. Existing COD methods primarily employ semantic segmentation, which suffers from overconfident incorrect predictions. In this paper, we propose a new paradigm that treats COD as a conditional mask-generation task leveraging diffusion models. Our method, dubbed CamoDiffusion, employs the denoising process of diffusion models to iteratively reduce the noise of the mask. Due to the stochastic sampling process of diffusion, our model is capable of sampling multiple possible predictions from the mask distribution, avoiding the problem of overconfident point estimation. Moreover, we develop specialized learning strategies that include an innovative ensemble approach for generating robust predictions and tailored forward diffusion methods for efficient training, specifically for the COD task. Extensive experiments on three COD datasets attest the superior performance of our model compared to existing state-of-the-art methods, particularly on the most challenging COD10K dataset, where our approach achieves 0.019 in terms of MAE.
Handwritten Mathematical Expression Recognition (HMER) is a challenging task with many educational applications. Recent methods for HMER have been developed for complex mathematical expressions in standard horizontal format. However, solutions for elementary mathematical expression, such as vertical addition and subtraction, have not been explored in the literature. This work proposes a new handwritten elementary mathematical expression dataset composed of addition and subtraction expressions in a vertical format. We also extended the MNIST dataset to generate artificial images with this structure. Furthermore, we proposed a solution for offline HMER, able to recognize vertical addition and subtraction expressions. Our analysis evaluated the object detection algorithms YOLO v7, YOLO v8, YOLO-NAS, NanoDet and FCOS for identifying the mathematical symbols. We also proposed a transcription method to map the bounding boxes from the object detection stage to a mathematical expression in the LATEX markup sequence. Results show that our approach is efficient, achieving a high expression recognition rate. The code and dataset are available at https://github.com/Danielgol/HME-VAS
Prevalent human-object interaction (HOI) detection approaches typically leverage large-scale visual-linguistic models to help recognize events involving humans and objects. Though promising, models trained via contrastive learning on text-image pairs often neglect mid/low-level visual cues and struggle at compositional reasoning. In response, we introduce DIFFUSIONHOI, a new HOI detector shedding light on text-to-image diffusion models. Unlike the aforementioned models, diffusion models excel in discerning mid/low-level visual concepts as generative models, and possess strong compositionality to handle novel concepts expressed in text inputs. Considering diffusion models usually emphasize instance objects, we first devise an inversion-based strategy to learn the expression of relation patterns between humans and objects in embedding space. These learned relation embeddings then serve as textual prompts, to steer diffusion models generate images that depict specific interactions, and extract HOI-relevant cues from images without heavy fine-tuning. Benefited from above, DIFFUSIONHOI achieves SOTA performance on three datasets under both regular and zero-shot setups.
Field of Parallax (FoP)}, a spatial field that distills the common features from different LF representations to provide flexible and consistent support for multi-task learning. FoP is built upon three core features--projection difference, adjacency divergence, and contextual consistency--which are essential for cross-task adaptability. To implement FoP, we design a two-step angular adapter: the first step captures angular-specific differences, while the second step consolidates contextual consistency to ensure robust representation. Leveraging the FoP-based representation, we introduce the LFX framework, the first to handle arbitrary LF representations seamlessly, unifying LF multi-task vision. We evaluated LFX across three different tasks, achieving new state-of-the-art results, compared with previous task-specific architectures: 84.74% in mIoU for semantic segmentation on UrbanLF, 0.84% in AP for object detection on PKU, and 0.030 in MAE and 0.026 in MAE for salient object detection on Duftv2 and PKU, respectively. The source code will be made publicly available at https://github.com/warriordby/LFX.
RGB-열화상(RGB-T) 객체 검출기는 가시광 영상과 열화상을 결합하여 강건한 검출을 수행하지만, 어느 한 스트림의 신호 손상이라도 정확도를 크게 저하시킬 수 있다. 복원 기반 방법은 상당한 지연을 초래하며 하위 정확도를 일관되게 향상시키지도 못하므로, 보다 직접적인 질문이 제기된다. 손상된 입력을 복원하는 대신, 변화하는 신호 품질에 검출기를 효율적으로 적응시킬 수 있는가? 기존 PEFT는 효율적인 경로를 제공하지만 입력에 독립적인 보정으로는 다양한 테스트 시점의 손상을 추적할 수 없다. 통제된 손상 변화 실험에 따르면 손상 유형은 깨끗한 조건 근처에서는 유사한 영향을 미치지만 심각도가 커질수록 점점 갈라지며, 이는 손상 인식형 입력별 적응의 동기가 된다. 이에 우리는 각 모달리티의 추정된 손상 유형과 심각도에 RGB-T 객체 검출기를 적응시키는 PG-FM(Prompt-Guided Feature Modulation)을 제안한다. 통제된 RGB-T 손상 시뮬레이션에서 PG-FM은 깨끗한 데이터셋 성능을 보존하고, FLIR에서 mAP50을 최대 20.8포인트, KAIST에서 최대 19.9포인트 향상시키며, 재학습 없이 수정되지 않은 자연 취득 외부 RGB-T 데이터셋으로 전이된다.
Deep learning has established many new state of the art solutions in the last decade in areas such as object, scene and speech recognition. In particular Convolutional Neural Network (CNN) is a category of deep learning which obtains excellent results in object detection and recognition tasks. Its architecture is indeed well suited to object analysis by learning and classifying complex (deep) features that represent parts of an object or the object itself. However, some of its features are very similar to texture analysis methods. CNN layers can be thought of as filter banks of complexity increasing with the depth. Filter banks are powerful tools to extract texture features and have been widely used in texture analysis. In this paper we develop a simple network architecture named Texture CNN (T-CNN) which explores this observation. It is built on the idea that the overall shape information extracted by the fully connected layers of a classic CNN is of minor importance in texture analysis. Therefore, we pool an energy measure from the last convolution layer which we connect to a fully connected layer. We show that our approach can improve the performance of a network while greatly reducing the memory usage and computation.
In the CMS experiment at CERN, Geneva, a large number of HGCAL sensor modules are fabricated in advanced laboratories around the world. Each sensor module contains about 700 checkpoints for visual inspection thus making it almost impossible to carry out such inspection manually. As artificial intelligence is more and more widely used in manufacturing, traditional detection technologies are gradually being intelligent. In order to more accurately evaluate the checkpoints, we propose to use deep learning-based object detection techniques to detect manufacturing defects in testing large numbers of modules automatically.
Segmenting or detecting objects in sparse Lidar point clouds are two important tasks in autonomous driving to allow a vehicle to act safely in its 3D environment. The best performing methods in 3D semantic segmentation or object detection rely on a large amount of annotated data. Yet annotating 3D Lidar data for these tasks is tedious and costly. In this context, we propose a self-supervised pre-training method for 3D perception models that is tailored to autonomous driving data. Specifically, we leverage the availability of synchronized and calibrated image and Lidar sensors in autonomous driving setups for distilling self-supervised pre-trained image representations into 3D models. Hence, our method does not require any point cloud nor image annotations. The key ingredient of our method is the use of superpixels which are used to pool 3D point features and 2D pixel features in visually similar regions. We then train a 3D network on the self-supervised task of matching these pooled point features with the corresponding pooled image pixel features. The advantages of contrasting regions obtained by superpixels are that: (1) grouping together pixels and points of visually coherent regions leads to a more meaningful contrastive task that produces features well adapted to 3D semantic segmentation and 3D object detection; (2) all the different regions have the same weight in the contrastive loss regardless of the number of 3D points sampled in these regions; (3) it mitigates the noise produced by incorrect matching of points and pixels due to occlusions between the different sensors. Extensive experiments on autonomous driving datasets demonstrate the ability of our image-to-Lidar distillation strategy to produce 3D representations that transfer well on semantic segmentation and object detection tasks.
The YOLO series has become the most popular framework for real-time object detection due to its reasonable trade-off between speed and accuracy. However, we observe that the speed and accuracy of YOLOs are negatively affected by the NMS. Recently, end-to-end Transformer-based detectors (DETRs) have provided an alternative to eliminating NMS. Nevertheless, the high computational cost limits their practicality and hinders them from fully exploiting the advantage of excluding NMS. In this paper, we propose the Real-Time DEtection TRansformer (RT-DETR), the first real-time end-to-end object detector to our best knowledge that addresses the above dilemma. We build RT-DETR in two steps, drawing on the advanced DETR: first we focus on maintaining accuracy while improving speed, followed by maintaining speed while improving accuracy. Specifically, we design an efficient hybrid encoder to expeditiously process multi-scale features by decoupling intra-scale interaction and cross-scale fusion to improve speed. Then, we propose the uncertainty-minimal query selection to provide high-quality initial queries to the decoder, thereby improving accuracy. In addition, RT-DETR supports flexible speed tuning by adjusting the number of decoder layers to adapt to various scenarios without retraining. Our RT-DETR-R50 / R101 achieves 53.1% / 54.3% AP on COCO and 108 / 74 FPS on T4 GPU, outperforming previously advanced YOLOs in both speed and accuracy. We also develop scaled RT-DETRs that outperform the lighter YOLO detectors (S and M models). Furthermore, RT-DETR-R50 outperforms DINO-R50 by 2.2% AP in accuracy and about 21 times in FPS. After pre-training with Objects365, RT-DETR-R50 / R101 achieves 55.3% / 56.2% AP. The project page: https://zhao-yian.github.io/RTDETR.
Conventional wheeled robots are unable to traverse scientifically interesting, but dangerous, cave environments. Multi-limbed climbing robot designs, such as ReachBot, are able to grasp irregular surface features and execute climbing motions to overcome obstacles, given suitable grasp locations. To support grasp site identification, we present a method for detecting rock cracks and edges, the SKeleton Intersection Loss (SKIL). SKIL is a loss designed for thin object segmentation that leverages the skeleton of the label. A dataset of rock face images was collected, manually annotated, and augmented with generated data. A new group of metrics, LineAcc, has been proposed for thin object segmentation such that the impact of the object width on the score is minimized. In addition, the metric is less sensitive to translation which can often lead to a score of zero when computing classical metrics such as Dice on thin objects. Our fine-tuned models outperform previous methods on similar thin object segmentation tasks such as blood vessel segmentation and show promise for integration onto a robotic system.
6D pose estimation of novel objects without CAD models remains an open challenge: model-free methods must build the object representation from scratch during onboarding and have long trailed their model-based counterparts. We present NovaSplat, a model-free pipeline that closes this gap. From a short smartphone video captured at onboarding, we train a 3D Gaussian Splatting (3DGS) model that serves both as a template source for detection and coarse pose estimation and as a differentiable backbone for pose refinement. Our refiner optimizes the object-to-camera pose with combined photometric and geometric losses under a continuous rotation parametrization free of gradient singularities. On a household-object benchmark, NovaSplat surpasses all model-based methods. On a challenging industrial benchmark with textureless and specular surfaces, it outperforms all model-free competitors while substantially narrowing the gap to model-based ones. We further introduce NovaSplat-Lite, an inference variant that is $3.1\times$ faster at negligible accuracy cost. Code will be made available at https://github.com/POSE-Lab/novasplat .
본 연구의 주된 목표는 혼잡한 장면에서의 비디오 객체 카운팅(video object counting)을 발전시키는 것으로, 주어진 텍스트 또는 시각적 프롬프트(prompt)를 바탕으로 대상 범주의 모든 인스턴스를 강건하게 계수하는 것을 지향한다. 기존 방법들은 RGB 정보에만 의존하여 혼잡하고 가려짐(occlusion)이 존재하는 환경에서 판별 능력이 제한적이다. 이를 해결하기 위해, 본 논문에서는 깊이 유도 검출기(Depth-Guided Detector, DG-Det)와 범용 후처리 파이프라인을 제안한다. 멀티스케일 RGB-D 크로스 어텐션(cross-attention) 및 명시적 가려짐 예측과 깊이 단서를 통합함으로써, 제안하는 방법은 공간적 이해를 향상시키고 혼잡 및 가려진 장면에서 강건한 검출을 달성한다. 나아가, 프레임 간 중복 계수를 제거하기 위한 통합 중복 제거(de-duplication) 프레임워크를 도입한다. 향후 연구를 촉진하기 위해 깊이 정보와 시퀀스당 다중 객체 범주를 포함하는 새로운 RGB-D 비디오 객체 카운팅 데이터셋도 함께 공개한다. 광범위한 실험을 통해 제안 방법이 기존 베이스라인 대비 MAE를 62.01% 감소시키고 RMSE에서도 일관된 성능 향상을 달성함을 입증한다. 소스 코드는 https://github.com/streamer-AP/DG-Net 에서, 데이터셋은 https://huggingface.co/datasets/aerospace123/RGBD-VideoCount 에서 제공된다.
6-DoF grasp pose detection of multi-grasp and multi-object is a challenge task in the field of intelligent robot. To imitate human reasoning ability for grasping objects, data driven methods are widely studied. With the introduction of large-scale datasets, we discover that a single physical metric usually generates several discrete levels of grasp confidence scores, which cannot finely distinguish millions of grasp poses and leads to inaccurate prediction results. In this paper, we propose a hybrid physical metric to solve this evaluation insufficiency. First, we define a novel metric is based on the force-closure metric, supplemented by the measurement of the object flatness, gravity and collision. Second, we leverage this hybrid physical metric to generate elaborate confidence scores. Third, to learn the new confidence scores effectively, we design a multi-resolution network called Flatness Gravity Collision GraspNet (FGC-GraspNet). FGC-GraspNet proposes a multi-resolution features learning architecture for multiple tasks and introduces a new joint loss function that enhances the average precision of the grasp detection. The network evaluation and adequate real robot experiments demonstrate the effectiveness of our hybrid physical metric and FGC-GraspNet. Our method achieves 90.5\% success rate in real-world cluttered scenes. Our code is available at https://github.com/luyh20/FGC-GraspNet.
In machine learning, Neural Architecture Search (NAS) requires domain knowledge of model design and a large amount of trial-and-error to achieve promising performance. Meanwhile, evolutionary algorithms have traditionally relied on fixed rules and pre-defined building blocks. The Large Language Model (LLM)-Guided Evolution (GE) framework transformed this approach by incorporating LLMs to directly modify model source code for image classification algorithms on CIFAR data and intelligently guide mutations and crossovers. A key element of LLM-GE is the "Evolution of Thought" (EoT) technique, which establishes feedback loops, allowing LLMs to refine their decisions iteratively based on how previous operations performed. In this study, we perform NAS for object detection by improving LLM-GE to modify the architecture of You Only Look Once (YOLO) models to enhance performance on the KITTI dataset. Our approach intelligently adjusts the design and settings of YOLO to find the optimal algorithms against objective such as detection accuracy and speed. We show that LLM-GE produced variants with significant performance improvements, such as an increase in Mean Average Precision from 92.5% to 94.5%. This result highlights the flexibility and effectiveness of LLM-GE on real-world challenges, offering a novel paradigm for automated machine learning that combines LLM-driven reasoning with evolutionary strategies.
This report describes a 2nd place solution of the detection challenge which is held within CVPR 2020 Retail-Vision workshop. Instead of going further considering previous results this work mainly aims to verify previously observed takeaways by re-experimenting. The reliability and reproducibility of the results are reached by incorporating a popular object detection toolbox - MMDetection. In this report, I firstly represent the results received for Faster-RCNN and RetinaNet models, which were taken for comparison in the original work. Then I describe the experiment results with more advanced models. The final section reviews two simple tricks for Faster-RCNN model that were used for my final submission: changing default anchor scale parameter and train-time image tiling. The source code is available at https://github.com/tyomj/product_detection.
The integrity of offshore wind turbines, pivotal for sustainable energy generation, is often compromised by surface material defects. Despite the availability of various detection techniques, limitations persist regarding cost-effectiveness, efficiency, and applicability. Addressing these shortcomings, this study introduces a novel approach leveraging an advanced version of the YOLOv8 object detection model, supplemented with a Convolutional Block Attention Module (CBAM) for improved feature recognition. The optimized loss function further refines the learning process. Employing a dataset of 5,432 images from the Saemangeum offshore wind farm and a publicly available dataset, our method underwent rigorous testing. The findings reveal a substantial enhancement in defect detection stability, marking a significant stride towards efficient turbine maintenance. This study's contributions illuminate the path for future research, potentially revolutionizing sustainable energy practices.
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.