Bounding Box Annotations Research Articles

GooseDetectlion: A Fully Annotated Dataset for Lion-head Goose Detection in Smart Farms

Large datasets are required to develop Artificial Intelligence (AI) models in AI powered smart farming for reducing farmers’ routine workload, this paper contributes the first large lion-head goose dataset GooseDetectlion, which consists of 2,660 images and 98,111 bounding box annotations. The dataset was collected with 6 cameras deployed in a goose farm in Chenghai district of Shantou city, Guangdong province, China. Images sampled from videos collected during July 9 -10 in 2022 were fully annotated by a team of fifty volunteers. Compared with another 6 well known animal datasets in literature, our dataset has higher capacity and density, which provides a challenging detection benchmark for main stream object detectors. Six state-of-the-art object detectors have been selected to be evaluated on the GooseDetectlion, which includes one two-stage anchor-based detector, three one-stage anchor-based detectors, as well as two one-stage anchor-free detectors. The results suggest that the one-stage anchor-based detector You Only Look Once version 5 (YOLO v5) achieves the best overall performance in terms of detection precision, model size and inference efficiency.

Detection and Classification of Cannabis Seeds Using RetinaNet and Faster R-CNN

The rapid growth of the cannabis industry necessitates accurate and efficient methods for detecting and classifying cannabis seed varieties, which is crucial for quality control, regulatory compliance, and genetic research. This study presents a deep learning approach to automate the detection and classification of 17 different cannabis seed varieties, addressing the limitations of manual inspection processes. Leveraging a unique dataset of 3319 high-resolution seed images, we employ self-supervised bounding box annotation using the Grounding DINO model. Our research evaluates two prominent object detection models, Faster R-CNN and RetinaNet, with different backbone architectures (ResNet50, ResNet101, and ResNeXt101). Extensive experiments reveal that RetinaNet with a ResNet101 backbone achieves the highest strict mean average precision (mAP) of 0.9458 at IoU 0.5–0.95. At the same time, Faster R-CNN with ResNet50 excels at the relaxed 0.5 IoU threshold (0.9428 mAP) and maintains superior recall. Notably, the ResNeXt101 backbone, despite its complexity, shows slightly lower performance across most metrics than ResNet architectures. In terms of inference speed, the Faster R-CNN with a ResNeXt101 backbone demonstrates the fastest processing at 17.5 frames per second. This comprehensive evaluation, including performance-speed trade-offs and per-class detection analysis, highlights the potential of deep learning for automating cannabis seed analysis. Our findings address challenges in seed purity, consistency, and regulatory adherence within the cannabis agricultural domain, paving the way for improved productivity and quality control in the industry.

DART: An automated end-to-end object detection pipeline with data Diversification, open-vocabulary bounding box Annotation, pseudo-label Review, and model Training

Accurate real-time object detection is vital across numerous industrial applications, from safety monitoring to quality control. Traditional approaches, however, are hindered by arduous manual annotation and data collection, struggling to adapt to ever-changing environments and novel target objects. To address these limitations, this paper presents DART, an innovative automated end-to-end pipeline that revolutionizes object detection workflows from data collection to model evaluation. It eliminates the need for laborious human labeling and extensive data collection while achieving outstanding accuracy across diverse scenarios. DART encompasses four key stages: (1) Data Diversification using subject-driven image generation (DreamBooth with SDXL), (2) Annotation via open-vocabulary object detection (Grounding DINO) to generate bounding box and class labels, (3) Review of generated images and pseudo-labels by large multimodal models (InternVL-1.5 and GPT-4o) to guarantee credibility, and (4) Training of real-time object detectors (YOLOv8 and YOLOv10) using the verified data. We apply DART to a self-collected dataset of construction machines named Liebherr Product, which contains over 15K high-quality images across 23 categories. The current instantiation of DART significantly increases average precision (AP) from 0.064 to 0.832. Its modular design ensures easy exchangeability and extensibility, allowing for future algorithm upgrades, seamless integration of new object categories, and adaptability to customized environments without manual labeling and additional data collection. The code and dataset are released at https://github.com/chen-xin-94/DART.

Challenges with segmenting intraoperative ultrasound for brain tumours

Objective - Addressing the challenges that come with identifying and delineating brain tumours in intraoperative ultrasound. Our goal is to both qualitatively and quantitatively assess the interobserver variation, amongst experienced neuro-oncological intraoperative ultrasound users (neurosurgeons and neuroradiologists), in detecting and segmenting brain tumours on ultrasound. We then propose that, due to the inherent challenges of this task, annotation by localisation of the entire tumour mass with a bounding box could serve as an ancillary solution to segmentation for clinical training, encompassing margin uncertainty and the curation of large datasets. Methods - 30 ultrasound images of brain lesions in 30 patients were annotated by 4 annotators - 1 neuroradiologist and 3 neurosurgeons. The annotation variation of the 3 neurosurgeons was first measured, and then the annotations of each neurosurgeon were individually compared to the neuroradiologist’s, which served as a reference standard as their segmentations were further refined by cross-reference to the preoperative magnetic resonance imaging (MRI). The following statistical metrics were used: Intersection Over Union (IoU), Sørensen-Dice Similarity Coefficient (DSC) and Hausdorff Distance (HD). These annotations were then converted into bounding boxes for the same evaluation. Results - There was a moderate level of interobserver variance between the neurosurgeons [IoU:0.789,DSC:0.876,HD:103.227]\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$[IoU:0.789, \\ DSC:0.876, \\ HD:103.227]$$\\end{document} and a larger level of variance when compared against the MRI-informed reference standard annotations by the neuroradiologist, mean across annotators [IoU:0.723,DSC:0.813,HD:115.675]\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$[IoU:0.723, \\ DSC:0.813, \\ HD:115.675]$$\\end{document}. After converting the segments to bounding boxes, all metrics improve, most significantly, the interquartile range drops by [IoU:37%,DSC:41%,HD:54%]\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$[IoU:37\\%, \\ DSC:41\\%, \\ HD:54\\%]$$\\end{document}. Conclusion - This study highlights the current challenges with detecting and defining tumour boundaries in neuro-oncological intraoperative brain ultrasound. We then show that bounding box annotation could serve as a useful complementary approach for both clinical and technical reasons.

ECC-PolypDet: Enhanced CenterNet With Contrastive Learning for Automatic Polyp Detection.

Accurate polyp detection is critical for early colorectal cancer diagnosis. Although remarkable progress has been achieved in recent years, the complex colon environment and concealed polyps with unclear boundaries still pose severe challenges in this area. Existing methods either involve computationally expensive context aggregation or lack prior modeling of polyps, resulting in poor performance in challenging cases. In this paper, we propose the Enhanced CenterNet with Contrastive Learning (ECC-PolypDet), a two-stage training & end-to-end inference framework that leverages images and bounding box annotations to train a general model and fine-tune it based on the inference score to obtain a final robust model. Specifically, we conduct Box-assisted Contrastive Learning (BCL) during training to minimize the intra-class difference and maximize the inter-class difference between foreground polyps and backgrounds, enabling our model to capture concealed polyps. Moreover, to enhance the recognition of small polyps, we design the Semantic Flow-guided Feature Pyramid Network (SFFPN) to aggregate multi-scale features and the Heatmap Propagation (HP) module to boost the model's attention on polyp targets. In the fine-tuning stage, we introduce the IoU-guided Sample Re-weighting (ISR) mechanism to prioritize hard samples by adaptively adjusting the loss weight for each sample during fine-tuning. Extensive experiments on six large-scale colonoscopy datasets demonstrate the superiority of our model compared with previous state-of-the-art detectors.

Implementation of surface crack detection method for nuclear fuel pellets by weakly supervised learning

ABSTRACT Surface cracks are one of the primary defects in nuclear fuel pellets, posing a significant hazard to nuclear safety production. Deep learning-based methods recently developed for crack detection are typically trained using a supervised learning, with detection performance dependent on pixel-level annotations, which suffer from high labeling costs and low efficiency. To address this issue, we propose a Weakly Supervised Crack Detection (WSCD) network for surface crack detection of nuclear fuel pellets. The method adopts bounding-box annotations instead of pixel-level annotations, and rapidly generates pixel-level pseudo-labels using the designed Local Fusion Segmentation (LFS) module. Leveraging the Mask-RCNN network as the backbone, the network introduces spatial attention mechanisms to optimize feature extraction networks, enhancing the extraction capability of multi-scale crack features. Lastly, a novel loss function is optimized to address sample imbalance issues and expedite network convergence. Experimental results on the established crack dataset demonstrate that the proposed method improves labeling efficiency by approximately 20 times, achieving a segmentation accuracy (IoU) of 81.9%, which reaches 92.2% of the segmentation accuracy achieved by supervised learning and outperforms that the IoU (80.3%) of other supervised crack segmentation network for nuclear fuel pellets, and meeting the precision requirements of nuclear fuel pellets production lines.

Quantification of cardiac capillarization in basement-membrane-immunostained myocardial slices using Segment Anything Model

Decreased myocardial capillary density has been reported as an important histopathological feature associated with various heart disorders. Quantitative assessment of cardiac capillarization typically involves double immunostaining of cardiomyocytes (CMs) and capillaries in myocardial slices. In contrast, single immunostaining of basement membrane protein is a straightforward approach to simultaneously label CMs and capillaries, presenting fewer challenges in background staining. However, subsequent image analysis always requires expertise and laborious manual work to identify and segment CMs/capillaries. Here, we developed an image analysis tool, AutoQC, for automatic identification and segmentation of CMs and capillaries in immunofluorescence images of basement membrane. Commonly used capillarization-related measurements can be derived from segmentation results. By leveraging the power of a pre-trained segmentation model (Segment Anything Model, SAM) via prompt engineering, the training of AutoQC required only a small dataset with bounding box annotations instead of pixel-wise annotations. AutoQC outperformed SAM (without prompt engineering) and YOLOv8-Seg, a state-of-the-art instance segmentation model, in both instance segmentation and capillarization assessment. Thus, AutoQC, featuring a weakly supervised algorithm, enables automatic segmentation and high-throughput, high-accuracy capillarization assessment in basement-membrane-immunostained myocardial slices. This approach reduces the training workload and eliminates the need for manual image analysis once AutoQC is trained.

Box2Mask: Box-Supervised Instance Segmentation via Level-Set Evolution.

In contrast to fully supervised methods using pixel-wise mask labels, box-supervised instance segmentation takes advantage of simple box annotations, which has recently attracted increasing research attention. This paper presents a novel single-shot instance segmentation approach, namely Box2Mask, which integrates the classical level-set evolution model into deep neural network learning to achieve accurate mask prediction with only bounding box supervision. Specifically, both the input image and its deep features are employed to evolve the level-set curves implicitly, and a local consistency module based on a pixel affinity kernel is used to mine the local context and spatial relations. Two types of single-stage frameworks, i.e., CNN-based and transformer-based frameworks, are developed to empower the level-set evolution for box-supervised instance segmentation, and each framework consists of three essential components: instance-aware decoder, box-level matching assignment and level-set evolution. By minimizing the level-set energy function, the mask map of each instance can be iteratively optimized within its bounding box annotation. The experimental results on five challenging testbeds, covering general scenes, remote sensing, medical and scene text images, demonstrate the outstanding performance of our proposed Box2Mask approach for box-supervised instance segmentation. In particular, with the Swin-Transformer large backbone, our Box2Mask obtains 42.4% mask AP on COCO, which is on par with the recently developed fully mask-supervised methods.

ChoreoVis: Planning and Assessing Formations in Dance Choreographies

AbstractSports visualization has developed into an active research field over the last decades. Many approaches focus on analyzing movement data recorded from unstructured situations, such as soccer. For the analysis of choreographed activities like formation dancing, however, the goal differs, as dancers follow specific formations in coordinated movement trajectories. To date, little work exists on how visual analytics methods can support such choreographed performances. To fill this gap, we introduce a new visual approach for planning and assessing dance choreographies. In terms of planning choreographies, we contribute a web application with interactive authoring tools and views for the dancers' positions and orientations, movement trajectories, poses, dance floor utilization, and movement distances. For assessing dancers' real‐world movement trajectories, extracted by manual bounding box annotations, we developed a timeline showing aggregated trajectory deviations and a dance floor view for detailed trajectory comparison. Our approach was developed and evaluated in collaboration with dance instructors, showing that introducing visual analytics into this domain promises improvements in training efficiency for the future.

Human behaviour detection dataset (HBDset) using computer vision for evacuation safety and emergency management

During emergency evacuation, it is crucial to accurately detect and classify different groups of evacuees based on their behaviours using computer vision. Traditional object detection models trained on standard image databases often fail to recognise individuals in specific groups such as the elderly, disabled individuals and pregnant women, who require additional assistance during emergencies. To address this limitation, this study proposes a novel image dataset called the Human Behaviour Detection Dataset (HBDset), specifically collected and annotated for public safety and emergency response purposes. This dataset contains eight types of human behaviour categories, i.e. the normal adult, child, holding a crutch, holding a baby, using a wheelchair, pregnant woman, lugging luggage and using a mobile phone. The dataset comprises more than 1,500 images collected from various public scenarios, with more than 2,900 bounding box annotations. The images were carefully selected, cleaned and subsequently manually annotated using the LabelImg tool. To demonstrate the effectiveness of the dataset, classical object detection algorithms were trained and tested based on the HBDset, and the average detection accuracy exceeds 90 %, highlighting the robustness and universality of the dataset. The developed open HBDset has the potential to enhance public safety, provide early disaster warnings and prioritise the needs of vulnerable individuals during emergency evacuation.

REAL-Colon: A dataset for developing real-world AI applications in colonoscopy

Detection and diagnosis of colon polyps are key to preventing colorectal cancer. Recent evidence suggests that AI-based computer-aided detection (CADe) and computer-aided diagnosis (CADx) systems can enhance endoscopists' performance and boost colonoscopy effectiveness. However, most available public datasets primarily consist of still images or video clips, often at a down-sampled resolution, and do not accurately represent real-world colonoscopy procedures. We introduce the REAL-Colon (Real-world multi-center Endoscopy Annotated video Library) dataset: a compilation of 2.7 M native video frames from sixty full-resolution, real-world colonoscopy recordings across multiple centers. The dataset contains 350k bounding-box annotations, each created under the supervision of expert gastroenterologists. Comprehensive patient clinical data, colonoscopy acquisition information, and polyp histopathological information are also included in each video. With its unprecedented size, quality, and heterogeneity, the REAL-Colon dataset is a unique resource for researchers and developers aiming to advance AI research in colonoscopy. Its openness and transparency facilitate rigorous and reproducible research, fostering the development and benchmarking of more accurate and reliable colonoscopy-related algorithms and models.

Self-supervised 3D vehicle detection based on monocular images

The deep learning-based 3D object detection literature on monocular images has been dominated by methods that require supervision in the form of 3D bounding box annotations for training. However, obtaining sufficient 3D annotations is expensive, laborious and prone to introducing errors. To address this problem, we propose a monocular self-supervised approach towards 3D object detection relying solely on observed RGB data rather than 3D bounding boxes for training. We leverage differentiable rendering to apply visual alignment to depth maps, instance masks and point clouds for self-supervision. Furthermore, considering the complexity of autonomous driving scenes, we introduce a point cloud filter to reduce noise impact and design an automatic training set pruning strategy suitable for the self-supervised framework to further improve network performance. We provide detailed experiments on the KITTI benchmark and achieve competitive performance with existing self-supervised methods as well as some fully supervised methods.

Point based weakly semi-supervised biomarker detection with cross-scale and label assignment in retinal OCT images

Background and ObjectiveOptical coherence tomography (OCT) is currently one of the most advanced retinal imaging methods. Retinal biomarkers in OCT images are of clinical significance and can assist ophthalmologists in diagnosing lesions. Compared with fundus images, OCT can provide higher resolution segmentation. However, image annotation at the bounding box level needs to be performed by ophthalmologists carefully and is difficult to obtain. In addition, the large variation in shape of different retinal markers and the inconspicuous appearance of biomarkers make it difficult for existing deep learning-based methods to effectively detect them. To overcome the above challenges, we propose a novel network for the detection of retinal biomarkers in OCT images. MethodsWe first address the issue of labeling cost using a novel weakly semi-supervised object detection method with point annotations which can reduce bounding box-level annotation efforts. To extend the method to the detection of biomarkers in OCT images, we propose multiple consistent regularizations for point-to-box regression network to deal with the shortage of supervision, which aims to learn more accurate regression mappings. Furthermore, in the subsequent fully supervised detection, we propose a cross-scale feature enhancement module to alleviate the detection problems caused by the large-scale variation of biomarkers. We also propose a dynamic label assignment strategy to distinguish samples of different importance more flexibly, thereby reducing detection errors due to the indistinguishable appearance of the biomarkers. ResultsWhen using our detection network, our regressor also achieves an AP value of 20.83 s when utilizing a 5 % fully labeled dataset partition, surpassing the performance of other comparative methods at 5 % and 10 %. Even coming close to the 20.87 % result achieved by Point DETR under 20 % full labeling conditions. When using Group R-CNN as the point-to-box regressor, our detector achieves 27.21 % AP in the 50 % fully labeled dataset experiment. 7.42 % AP improvement is achieved compared to our detection network baseline Faster R-CNN. ConclusionsThe experimental findings not only demonstrate the effectiveness of our approach with minimal bounding box annotations but also highlight the enhanced biomarker detection performance of the proposed module. We have included a detailed algorithmic flow in the supplementary material.

Deep Omni-Supervised Learning for Rib Fracture Detection From Chest Radiology Images.

Deep learning (DL)-based rib fracture detection has shown promise of playing an important role in preventing mortality and improving patient outcome. Normally, developing DL-based object detection models requires a huge amount of bounding box annotation. However, annotating medical data is time-consuming and expertise-demanding, making obtaining a large amount of fine-grained annotations extremely infeasible. This poses a pressing need for developing label-efficient detection models to alleviate radiologists' labeling burden. To tackle this challenge, the literature on object detection has witnessed an increase of weakly-supervised and semi-supervised approaches, yet still lacks a unified framework that leverages various forms of fully-labeled, weakly-labeled, and unlabeled data. In this paper, we present a novel omni-supervised object detection network, ORF-Netv2, to leverage as much available supervision as possible. Specifically, a multi-branch omni-supervised detection head is introduced with each branch trained with a specific type of supervision. A co-training-based dynamic label assignment strategy is then proposed to enable flexible and robust learning from the weakly-labeled and unlabeled data. Extensive evaluation was conducted for the proposed framework with three rib fracture datasets on both chest CT and X-ray. By leveraging all forms of supervision, ORF-Netv2 achieves mAPs of 34.7, 44.7, and 19.4 on the three datasets, respectively, surpassing the baseline detector which uses only box annotations by mAP gains of 3.8, 4.8, and 5.0, respectively. Furthermore, ORF-Netv2 consistently outperforms other competitive label-efficient methods over various scenarios, showing a promising framework for label-efficient fracture detection. The code is available at: https://github.com/zhizhongchai/ORF-Net.

Instrument Detection and Descriptive Gesture Segmentation on a Robotic Surgical Maneuvers Dataset

Large datasets play a crucial role in the progression of surgical robotics, facilitating advancements in the fields of surgical task recognition and automation. Moreover, public datasets enable the comparative analysis of various algorithms and methodologies, thereby assessing their effectiveness and performance. The ROSMA (Robotics Surgical Maneuvers) dataset provides 206 trials of common surgical training tasks performed with the da Vinci Research Kit (dVRK). In this work, we extend the ROSMA dataset with two annotated subsets: ROSMAT24, which contains bounding box annotations for instrument detection, and ROSMAG40, which contains high and low-level gesture annotations. We propose an annotation method that provides independent labels for the right-handed tools and the left-handed tools. For instrument identification, we validate our proposal with a YOLOv4 model in two experimental scenarios. We demonstrate the generalization capabilities of the network to detect instruments in unseen scenarios. On the other hand, for gesture segmentation, we propose two label categories: high-level annotations that describe gestures at a maneuvers level, and low-level annotations that describe gestures at a fine-grain level. To validate this proposal, we have designed a recurrent neural network based on a bidirectional long-short term memory layer. We present results for four cross-validation experimental setups, reaching up to a 77.35% mAP.

On-the-fly point annotation for fast medical video labeling.

In medical research, deep learning models rely on high-quality annotated data, a process often laborious and time-consuming. This is particularly true for detection tasks where bounding box annotations are required. The need to adjust two corners makes the process inherently frame-by-frame. Given the scarcity of experts' time, efficient annotation methods suitable for clinicians are needed. We propose an on-the-fly method for live video annotation to enhance the annotation efficiency. In this approach, a continuous single-point annotation is maintained by keeping the cursor on the object in a live video, mitigating the need for tedious pausing and repetitive navigation inherent in traditional annotation methods. This novel annotation paradigm inherits the point annotation's ability to generate pseudo-labels using a point-to-box teacher model. We empirically evaluate this approach by developing a dataset and comparing on-the-fly annotation time against traditional annotation method. Using our method, annotation speed was faster than the traditional annotation technique. We achieved a mean improvement of AP@50 over conventional method at equivalent annotation budgets on the developed dataset. Without bells and whistles, our approach offers a significant speed-up in annotation tasks. It can be easily implemented on any annotation platform to accelerate the integration of deep learning in video-based medical research.

FNPC-SAM: Uncertainty-Guided False Negative/Positive Control for SAM on Noisy Medical Images.

The Segment Anything Model (SAM) is a recently developed all-range foundation model for image segmentation. It can use sparse manual prompts such as bounding boxes to generate pixel-level segmentation in natural images but struggles in medical images such as low-contrast, noisy ultrasound images. We propose a refined test-phase prompt augmentation technique designed to improve SAM's performance in medical image segmentation. The method couples multi-box prompt augmentation and an aleatoric uncertainty-based false-negative (FN) and false-positive (FP) correction (FNPC) strategy. We evaluate the method on two ultrasound datasets and show improvement in SAM's performance and robustness to inaccurate prompts, without the necessity for further training or tuning. Moreover, we present the Single-Slice-to-Volume (SS2V) method, enabling 3D pixel-level segmentation using only the bounding box annotation from a single 2D slice. Our results allow efficient use of SAM in even noisy, low-contrast medical images. The source code has been released at: https://github.com/MedICL-VU/FNPC-SAM.

Revisiting Computer-Aided Tuberculosis Diagnosis.

Tuberculosis (TB) is a major global health threat, causing millions of deaths annually. Although early diagnosis and treatment can greatly improve the chances of survival, it remains a major challenge, especially in developing countries. Recently, computer-aided tuberculosis diagnosis (CTD) using deep learning has shown promise, but progress is hindered by limited training data. To address this, we establish a large-scale dataset, namely the Tuberculosis X-ray (TBX11 K) dataset, which contains 11 200 chest X-ray (CXR) images with corresponding bounding box annotations for TB areas. This dataset enables the training of sophisticated detectors for high-quality CTD. Furthermore, we propose a strong baseline, SymFormer, for simultaneous CXR image classification and TB infection area detection. SymFormer incorporates Symmetric Search Attention (SymAttention) to tackle the bilateral symmetry property of CXR images for learning discriminative features. Since CXR images may not strictly adhere to the bilateral symmetry property, we also propose Symmetric Positional Encoding (SPE) to facilitate SymAttention through feature recalibration. To promote future research on CTD, we build a benchmark by introducing evaluation metrics, evaluating baseline models reformed from existing detectors, and running an online challenge. Experiments show that SymFormer achieves state-of-the-art performance on the TBX11 K dataset.

IDet3D: Towards Efficient Interactive Object Detection for LiDAR Point Clouds

Accurately annotating multiple 3D objects in LiDAR scenes is laborious and challenging. While a few previous studies have attempted to leverage semi-automatic methods for cost-effective bounding box annotation, such methods have limitations in efficiently handling numerous multi-class objects. To effectively accelerate 3D annotation pipelines, we propose iDet3D, an efficient interactive 3D object detector. Supporting a user-friendly 2D interface, which can ease the cognitive burden of exploring 3D space to provide click interactions, iDet3D enables users to annotate the entire objects in each scene with minimal interactions. Taking the sparse nature of 3D point clouds into account, we design a negative click simulation (NCS) to improve accuracy by reducing false-positive predictions. In addition, iDet3D incorporates two click propagation techniques to take full advantage of user interactions: (1) dense click guidance (DCG) for keeping user-provided information throughout the network and (2) spatial click propagation (SCP) for detecting other instances of the same class based on the user-specified objects. Through our extensive experiments, we present that our method can construct precise annotations in a few clicks, which shows the practicality as an efficient annotation tool for 3D object detection.

WeakPCSOD: Overcoming the Bias of Box Annotations for Weakly Supervised Point Cloud Salient Object Detection

Point cloud salient object detection (PCSOD) is a newly proposed task in 3D dense segmentation. However, the acquisition of accurate 3D dense annotations comes at a high cost, severely limiting the progress of PCSOD. To address this issue, we propose the first weakly supervised PCSOD (named WeakPCSOD) model, which relies solely on cheap 3D bounding box annotations. In WeakPCSOD, we extract noise-free supervision from coarse 3D bounding boxes while mitigating shape biases inherent in box annotations. To achieve this, we introduce a novel mask-to-box (M2B) transformation and a color consistency (CC) loss. The M2B transformation, from a shape perspective, disentangles predictions from labels, enabling the extraction of noiseless supervision from labels while preserving object shapes independently of the box bias. From an appearance perspective, we further introduce the CC loss to provide dense supervision, which mitigates the non-unique predictions stemming from weak supervision and substantially reduces prediction variability. Furthermore, we employ a self-training (ST) strategy to enhance performance by utilizing high-confidence pseudo labels. Notably, the M2B transformation, CC loss, and ST strategy are seamlessly integrated into any model and incur no computational costs for inference. Extensive experiments demonstrate the effectiveness of our WeakPCSOD model, even comparable to fully supervised models utilizing dense annotations.