3D Bounding Box Research Articles

Three-Dimensional Outdoor Object Detection in Quadrupedal Robots for Surveillance Navigations

Quadrupedal robots are confronted with the intricate challenge of navigating dynamic environments fraught with diverse and unpredictable scenarios. Effectively identifying and responding to obstacles is paramount for ensuring safe and reliable navigation. This paper introduces a pioneering method for 3D object detection, termed viewpoint feature histograms, which leverages the established paradigm of 2D detection in projection. By translating 2D bounding boxes into 3D object proposals, this approach not only enables the reuse of existing 2D detectors but also significantly increases the performance with less computation required, allowing for real-time detection. Our method is versatile, targeting both bird’s eye view objects (e.g., cars) and frontal view objects (e.g., pedestrians), accommodating various types of 2D object detectors. We showcase the efficacy of our approach through the integration of YOLO3D, utilizing LiDAR point clouds on the KITTI dataset, to achieve real-time efficiency aligned with the demands of autonomous vehicle navigation. Our model selection process, tailored to the specific needs of quadrupedal robots, emphasizes considerations such as model complexity, inference speed, and customization flexibility, achieving an accuracy of up to 99.93%. This research represents a significant advancement in enabling quadrupedal robots to navigate complex and dynamic environments with heightened precision and safety.

AEPF: Attention-Enabled Point Fusion for 3D Object Detection.

Current state-of-the-art (SOTA) LiDAR-only detectors perform well for 3D object detection tasks, but point cloud data are typically sparse and lacks semantic information. Detailed semantic information obtained from camera images can be added with existing LiDAR-based detectors to create a robust 3D detection pipeline. With two different data types, a major challenge in developing multi-modal sensor fusion networks is to achieve effective data fusion while managing computational resources. With separate 2D and 3D feature extraction backbones, feature fusion can become more challenging as these modes generate different gradients, leading to gradient conflicts and suboptimal convergence during network optimization. To this end, we propose a 3D object detection method, Attention-Enabled Point Fusion (AEPF). AEPF uses images and voxelized point cloud data as inputs and estimates the 3D bounding boxes of object locations as outputs. An attention mechanism is introduced to an existing feature fusion strategy to improve 3D detection accuracy and two variants are proposed. These two variants, AEPF-Small and AEPF-Large, address different needs. AEPF-Small, with a lightweight attention module and fewer parameters, offers fast inference. AEPF-Large, with a more complex attention module and increased parameters, provides higher accuracy than baseline models. Experimental results on the KITTI validation set show that AEPF-Small maintains SOTA 3D detection accuracy while inferencing at higher speeds. AEPF-Large achieves mean average precision scores of 91.13, 79.06, and 76.15 for the car class's easy, medium, and hard targets, respectively, in the KITTI validation set. Results from ablation experiments are also presented to support the choice of model architecture.

Six-Degree-of-Freedom Pose Estimation Method for Multi-Source Feature Points Based on Fully Convolutional Neural Network

An object’s six-degree-of-freedom (6DoF) pose information has great importance in various fields. Existing methods of pose estimation usually detect two-dimensional (2D)-three-dimensional (3D) feature point pairs, and directly estimates the pose information through Perspective-n-Point (PnP) algorithms. However, this approach ignores the spatial association between pixels, making it difficult to obtain high-precision results. In order to apply pose estimation based on deep learning methods to real-world scenarios, we hope to design a method that is robust enough in more complex scenarios. Therefore, we introduce a method for 3D object pose estimation from color images based on farthest point sampling (FPS) and object 3D bounding box. This method detects the 2D projection of 3D feature points through a convolutional neural network, matches it with the 3D model of the object, and then uses the PnP algorithm to restore the feature point pair to the object pose. Due to the global nature of the bounding box, this approach can be considered effective even in partially occluded or complex environments. In addition, we propose a heatmap suppression method based on weighted coordinates to further improve the prediction accuracy of feature points and the accuracy of the PnP algorithm in solving the pose position. Compared with other algorithms, this method has higher accuracy and better robustness. Our method yielded 93.8% of the ADD(-s) metrics on the Linemod dataset and 47.7% of the ADD(-s) metrics on the Occlusion Linemod dataset. These results show that our method is more effective than existing methods in pose estimation of large objects.

Enhanced 3D object detection for autonomous driving: A spatial-temporal alignment approach in Bird's Eye View scenarios

This paper presents a novel 3D object detection algorithm designed for Bird's Eye View (BEV) scenarios, which significantly improves detection capabilities by integrating spatial and temporal features. The core of our approach is the spatial-temporal alignment module that efficiently processes information across different time steps and spatial locations, enhancing the precision and robustness of object detection. We employ a temporal self-attention mechanism to capture the motion information of objects over time, allowing the model to correlate features across various time steps for identifying and tracking moving objects. Additionally, a spatial cross-attention mechanism is utilized to focus on spatial features within regions of interest, promoting interactions between features extracted from camera views and BEV queries. Our method also implements temporal feature integration and multi-scale feature fusion to enhance detection stability and accuracy for fast-moving objects and to capture multi-scale context information, respectively. The model employs an enriched feature set post alignment for 3D bounding box prediction, ascertaining the position, dimensions, and orientation of objects. We conducted experiments on two public datasets for autonomous driving nuScenes and Waymo Open Dataset, demonstrating that our method outperforms previous BEVFormer and other state-of-the-art methods in terms of detection accuracy and robustness. The paper concludes with potential future directions for optimizing the BEVFormer model's performance and exploring its application in broader scenarios and tasks.

Different radiomics annotation methods comparison in rectal cancer characterisation and prognosis prediction: a two-centre study

ObjectivesTo explore the performance differences of multiple annotations in radiomics analysis and provide a reference for tumour annotation in large-scale medical image analysis.MethodsA total of 342 patients from two centres who underwent radical resection for rectal cancer were retrospectively studied and divided into training, internal validation, and external validation cohorts. Three predictive tasks of tumour T-stage (pT), lymph node metastasis (pLNM), and disease-free survival (pDFS) were performed. Twelve radiomics models were constructed using Lasso-Logistic or Lasso-Cox to evaluate and four annotation methods, 2D detailed annotation along tumour boundaries (2D), 3D detailed annotation along tumour boundaries (3D), 2D bounding box (2DBB), and 3D bounding box (3DBB) on T2-weighted images, were compared. Radiomics models were used to establish combined models incorporating clinical risk factors. The DeLong test was performed to compare the performance of models using the receiver operating characteristic curves.ResultsFor radiomics models, the area under the curve values ranged from 0.627 (0.518–0.728) to 0.811 (0.705–0.917) in the internal validation cohort and from 0.619 (0.469–0.754) to 0.824 (0.689–0.918) in the external validation cohort. Most radiomics models based on four annotations did not differ significantly, except between the 3D and 3DBB models for pLNM (p = 0.0188) in the internal validation cohort. For combined models, only the 2D model significantly differed from the 2DBB (p = 0.0372) and 3D models (p = 0.0380) for pDFS.ConclusionRadiomics and combined models constructed with 2D and bounding box annotations showed comparable performances to those with 3D and detailed annotations along tumour boundaries in rectal cancer characterisation and prognosis prediction.Critical relevance statementFor quantitative analysis of radiological images, the selection of 2D maximum tumour area or bounding box annotation is as representative and easy to operate as 3D whole tumour or detailed annotations along tumour boundaries.Key PointsThere is currently a lack of discussion on whether different annotation efforts in radiomics are predictively representative.No significant differences were observed in radiomics and combined models regardless of the annotations (2D, 3D, detailed, or bounding box).Prioritise selecting the more time and effort-saving 2D maximum area bounding box annotation.Graphical

Weakly supervised detection of pheochromocytomas and paragangliomas in CT using noisy data

Pheochromocytomas and Paragangliomas (PPGLs) are rare adrenal and extra-adrenal tumors that have metastatic potential. Management of patients with PPGLs mainly depends on the makeup of their genetic cluster: SDHx, VHL/EPAS1, kinase, and sporadic. CT is the preferred modality for precise localization of PPGLs, such that their metastatic progression can be assessed. However, the variable size, morphology, and appearance of these tumors in different anatomical regions can pose challenges for radiologists. Since radiologists must routinely track changes across patient visits, manual annotation of PPGLs is quite time-consuming and cumbersome to do across all axial slices in a CT volume. As such, PPGLs are only weakly annotated on axial slices by radiologists in the form of RECIST measurements. To ameliorate the manual effort spent by radiologists, we propose a method for the automated detection of PPGLs in CT via a proxy segmentation task. Weak 3D annotations (derived from 2D bounding boxes) were used to train both 2D and 3D nnUNet models to detect PPGLs via segmentation. We evaluated our approaches on an in-house dataset comprised of chest-abdomen-pelvis CTs of 255 patients with confirmed PPGLs. On a test set of 53 CT volumes, our 3D nnUNet model achieved a detection precision of 70% and sensitivity of 64.1%, and outperformed the 2D model that obtained a precision of 52.7% and sensitivity of 27.5% (p< 0.05). SDHx and sporadic genetic clusters achieved the highest precisions of 73.1% and 72.7% respectively. Our state-of-the art findings highlight the promising nature of the challenging task of automated PPGL detection.

TIP-Editor: An Accurate 3D Editor Following Both Text-Prompts And Image-Prompts

Text-driven 3D scene editing has gained significant attention owing to its convenience and user-friendliness. However, existing methods still lack accurate control of the specified appearance and location of the editing result due to the inherent limitations of the text description. To this end, we propose a 3D scene editing framework, TIP-Editor, that accepts both text and image prompts and a 3D bounding box to specify the editing region. With the image prompt, users can conveniently specify the detailed appearance/style of the target content in complement to the text description, enabling accurate control of the appearance. Specifically, TIP-Editor employs a stepwise 2D personalization strategy to better learn the representation of the existing scene and the reference image, in which a localization loss is proposed to encourage correct object placement as specified by the bounding box. Additionally, TIP-Editor utilizes explicit and flexible 3D Gaussian splatting (GS) as the 3D representation to facilitate local editing while keeping the background unchanged. Extensive experiments have demonstrated that TIP-Editor conducts accurate editing following the text and image prompts in the specified bounding box region, consistently outperforming the baselines in editing quality, and the alignment to the prompts, qualitatively and quantitatively.

An improved LIO-SAM algorithm by integrating image information for dynamic and unstructured environments

Abstract Simultaneous localization and mapping (SLAM) is the process of estimating the trajectory of a mobile sensor carrier and creating a representation of its surroundings. Traditional SLAM algorithms are based on "static world assumption" and simplify the problem by filtering out moving objects or tracking them separately in complex dynamic environments. However, this strong assumption restricts the application of SLAM algorithms on highly dynamic and unstructured environments. In order to resolve above problem, this paper propose an improved object-aware dynamic SLAM system by integrating image information, i.e., semantic and velocity information. Firstly, we adopt deep learning method to detect both the 2D and 3D bounding boxes of objects in the environment. This information is then used to perform multi-view, multi-dimensional bundle optimization to jointly refine the poses of camera, object, and point. Secondly, 2D detection results from image and 3D detection results from lidar are integrated by the joint probabilistic data association (JPDA) data association algorithm to facilitate object-level data association. We also calculate 2D and 3D motion velocity and this information is used to constraint the motion of the object. Finally, we perform comprehensive experiments on different datasets, including NCLT, M2DGR, and KITTI to prove the performance of the proposed method.

An Advanced Approach to Object Detection and Tracking in Robotics and Autonomous Vehicles Using YOLOv8 and LiDAR Data Fusion

Accurately and reliably perceiving the environment is a major challenge in autonomous driving and robotics research. Traditional vision-based methods often suffer from varying lighting conditions, occlusions, and complex environments. This paper addresses these challenges by combining a deep learning-based object detection algorithm, YOLOv8, with LiDAR data fusion technology. The principle of this combination is to merge the advantages of these technologies: YOLOv8 excels in real-time object detection and classification through RGB images, while LiDAR provides accurate distance measurement and 3D spatial information, regardless of lighting conditions. The integration aims to apply the high accuracy and robustness of YOLOv8 in identifying and classifying objects, as well as the depth data provided by LiDAR. This combination enhances the overall environmental perception, which is critical for the reliability and safety of autonomous systems. However, this fusion brings some research challenges, including data calibration between different sensors, filtering ground points from LiDAR point clouds, and managing the computational complexity of processing large datasets. This paper presents a comprehensive approach to address these challenges. Firstly, a simple algorithm is introduced to filter out ground points from LiDAR point clouds, which are essential for accurate object detection, by setting different threshold heights based on the terrain. Secondly, YOLOv8, trained on a customized dataset, is utilized for object detection in images, generating 2D bounding boxes around detected objects. Thirdly, a calibration algorithm is developed to transform 3D LiDAR coordinates to image pixel coordinates, which are vital for correlating LiDAR data with image-based object detection results. Fourthly, a method for clustering different objects based on the fused data is proposed, followed by an object tracking algorithm to compute the 3D poses of objects and their relative distances from a robot. The Agilex Scout Mini robot, equipped with Velodyne 16-channel LiDAR and an Intel D435 camera, is employed for data collection and experimentation. Finally, the experimental results validate the effectiveness of the proposed algorithms and methods.

Promptable foundation model for automatic whole body RECIST measurement.

e13643 Background: RECIST 1.1 is the gold standard for tumor evaluation in clinical trials and patient care. However, it faces challenges due to the subjective selection and measurement of lesions, notably because of inter-observer variability in the 20-30% range (1), which can potentially result in inaccuracies in therapeutic response classification (2). The main objective of this study is to assess the added value of a radiological foundation model to assist the radiologist to measure RECIST across different lesion topographies, using visual prompting. Methods: We use a promptable segmentation algorithm, based on a visual foundation model, pre-trained on various CT-scans and then trained on a subset of the Medical Segmentation Decathlon (MSD) dataset, covering 538 patients with pancreas, liver, colon and lung lesions. It outputs a 3D segmentation mask of the lesion, using a visual prompt, i.e. a 3D bounding box (BBox) as input. We then measure the capacity of the model to measure RECIST on the MSD validation set and on the KiTS validation set (99 samples), which features a new lesion type (kidney). To simulate inter-reader variability of up to 23%, we provide as inputs BBoxes centered on the lesions but with a varying error (15% variability: +10 pixels, 23% variability: +15 pixels). Results: The model is able to correct variability in BBox input: with all BBox error ranges, it beats the input variability (15% and 23% thresholds in each column) across many organ types, except in the pancreas when using large error BBoxes. Moreover, the model is able to generalize to new lesion types not seen during training, on an external dataset (KiTS - Kidney). Conclusions: Unlike existing supervised machine learning models dedicated to lesion detections in specific organs, our approach is organ and lesion agnostic and offers a more reliable, precise tumor assessment. This is a first step before allowing the longitudinal evaluation of tumors while safeguarding the clinical intuition necessary for selecting the right target lesions. Moreover, we believe that these methods will be crucial in advancing beyond RECIST 1.1, facilitating the identification of new prognostic biomarkers and proxies for tumor burden derived from previously underexplored radiomics features, ultimately refining the efficacy of clinical trial assessments and elevating the standard of patient care. 1. Yoon, S et al. 2015. 2. Kuhl et al. 2018. [Table: see text]

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.

CEPB dataset: a photorealistic dataset to foster the research on bin picking in cluttered environments.

Several datasets have been proposed in the literature, focusing on object detection and pose estimation. The majority of them are interested in recognizing isolated objects or the pose of objects in well-organized scenarios. This work introduces a novel dataset that aims to stress vision algorithms in the difficult task of object detection and pose estimation in highly cluttered scenes concerning the specific case of bin picking for the Cluttered Environment Picking Benchmark (CEPB). The dataset provides about 1.5M virtually generated photo-realistic images (RGB + depth + normals + segmentation) of 50K annotated cluttered scenes mixing rigid, soft, and deformable objects of varying sizes used in existing robotic picking benchmarks together with their 3D models (40 objects). Such images include three different camera positions, three light conditions, and multiple High Dynamic Range Imaging (HDRI) maps for domain randomization purposes. The annotations contain the 2D and 3D bounding boxes of the involved objects, the centroids' poses (translation + quaternion), and the visibility percentage of the objects' surfaces. Nearly 10K separated object images are presented to perform simple tests and compare them with more complex cluttered scenarios tests. A baseline performed with the DOPE neural network is reported to highlight the challenges introduced by the novel dataset.

SRFDet3D: Sparse Region Fusion based 3D Object Detection

Unlike the earlier 3D object detection approaches that formulate hand-crafted dense (in thousands) object proposals by leveraging anchors on dense feature maps, we formulate np (in hundreds) number of learnable sparse object proposals to predict 3D bounding box parameters. The sparse proposals in our approach are not only learnt during training but also are input-dependent, so they represent better object candidates during inference. Leveraging the sparse proposals, we fuse only the sparse regions of multi-modal features and we propose Sparse Region Fusion based 3D object Detection (SRFDet3D) network with mainly three components: an encoder for feature extraction, a region proposal generation module for sparse input-dependent proposals and a decoder for multi-modal feature fusion and iterative refinement of object proposals. Additionally for optimal training, we formulate our sparse detector with many-to-one label assignment based on Optimal Transport Algorithm (OTA). We conduct extensive experiments and analysis on publicly available large-scale autonomous driving datasets: nuScenes, KITTI, and Waymo. Our LiDAR-only SRFDet3D-L network achieves 63.1 mAP and outperforms the state-of-the-art networks on the nuScenes dataset, surpassing the dense detectors on KITTI and Waymo datasets. Our LiDAR-Camera model SRFDet3D achieves 64.7 mAP with improvements over existing fusion methods.

A region feature fusion network for point cloud and image to detect 3D object

AbstractSensor fusion is very important for collaborative intelligent systems. A regional feature fusion network called ReFuNet for detecting 3D Object is proposed. It is difficult to detect distant or small objects accurately for the sparsity of LiDAR point cloud. The LiDAR point cloud and camera image information to solve the problem of point cloud sparsity is used, which can integrate image‐rich semantic information to enhance point cloud features. Also, the authors’ ReFuNet method segments the possible areas of objects by the results of 2D image detection. A cross‐attention mechanism adaptively fuses image and point cloud features within the areas. Then, the authors’ ReFuNet uses fused features to predict the 3D bounding boxes of objects. Experiments on the KITTI 3D object detection dataset showed that the authors’ proposed fusion method effectively improved the performance of 3D object detection.

MMoNet3D: Enhancing monocular 3D object localization with modern advancements

MMoNet3D, an innovative project, addresses monocular multi-object detection and 3D localization in real-time, surpassing the limitations of the existing MoNet3D system. Utilizing a custom deep learning model inspired by Centernet, it excels in oriented car detection across static images, videos, and live webcam feeds. The system, powered by the KITTI dataset, showcases its potential impact on real-time monocular multi-object detection. Its unique features, including 3D bounding box visualization and a user-friendly GUI, set MMoNet3D apart, making it suitable for embedded advanced driving-assistance systems. This advancement promises heightened accuracy and efficiency, contributing to enhanced vehicular safety and operational efficacy. MMoNet3D stands as a beacon in computer vision, paving the way for intelligent transportation systems.

YOLO2U-Net: Detection-guided 3D instance segmentation for microscopy

Microscopy imaging techniques are instrumental for characterization and analysis of biological structures. As these techniques typically render 3D visualization of cells by stacking 2D projections, issues such as out-of-plane excitation and low resolution in the z-axis may pose challenges (even for human experts) to detect individual cells in 3D volumes as these non-overlapping cells may appear as overlapping. A comprehensive method for accurate 3D instance segmentation of cells in the brain tissue is introduced here. The proposed method combines the 2D YOLO detection method with a multi-view fusion algorithm to construct a 3D localization of the cells. Next, the 3D bounding boxes along with the data volume are input to a 3D U-Net network that is designed to segment the primary cell in each 3D bounding box, and in turn, to carry out instance segmentation of cells in the entire volume. The promising performance of the proposed method is shown in comparison with current deep learning-based 3D instance segmentation methods.

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.

Geometric information constraint 3D object detection from LiDAR point cloud for autonomous vehicles under adverse weather

3D object detection, as the core of the autonomous vehicle perception module, is essential for efficient transportation and comfortable experiences. However, the challenge of 3D object detection under adverse weather conditions hinders the advancement of autonomous vehicles to higher levels. Hence, achieving accurate 3D object detection under adverse weather conditions is increasingly crucial as it forms the foundation for trajectory planning and driving strategy making in autonomous vehicles, thereby revolutionizing transportation modes for both goods and passengers. Advances in Light Detection and Ranging (LiDAR) technology have facilitated the development of 3D object detection in the past few years. Adverse weather, which inevitably occurs in real-world driving scenarios, could degrade measurement accuracy and point density of LiDAR and lead to particle interference. Detecting accurate 3D bounding boxes from sparse, incomplete point clouds with particle interference is difficult. Therefore, this research presents a novel geometric information constraint network for 3D object detection tasks from LiDAR point clouds under adverse weather (GIC-Net). In this study, we focus on how to incorporate geometric location information and line geometric feature information into the network against adverse weather. Further, we propose a geometric location constrained backbone module (GLC) to reduce rain and snow particle interference and ensure sufficient receptive fields. Then, we design a line geometric feature constraint module (LGFC) to add line constraints of 3D bounding boxes into the training process. Finally, a line loss function is designed, and features from the GLC and LGFC modules are fed into the multi-task detection head for accurate 3D bounding box prediction. Experiments on the Canadian Adverse Driving Conditions (CADC) autonomous vehicle dataset demonstrate the superiority of our method over six other state-of-the-art methods under adverse weather, which is at least 13.32 %, 4.67 %, and 10.44 % mAP higher than the other compared methods in the car, truck, and pedestrian classes respectively. Also, we further verify the better generalization ability of our network compared to other methods.

Enhancing pseudo label quality for pedestrian and cyclist in weakly supervised 3D object detection

Weakly supervised 3D object detection for autonomous driving primarily focuses on cars because of their distinct rectangle boundaries and abundant instances. However, detecting categories with ambiguous rectangle boundaries and fewer instances than cars, such as pedestrians and cyclists, remains challenging with limited research. Ambiguity in rectangle boundaries presents significant difficulties in generating accurate 3D pseudo labels, while the scarcity of instances often leads to convergence issues during detector training. Pedestrians and cyclists are dense inside the 3D bounding boxes but sparse at corners and boundaries. Density is a practical clue for locating and discriminating pedestrians and cyclists in point clouds. This paper proposes a density-based 3D pseudo-label generation module (DPL-3D), addressing the challenges of ambiguous rectangle boundaries. Ambiguity rectangle boundaries will lead to poor pseudo-label quality. Therefore, By leveraging the density information of 3D points, our DPL-3D improves the accuracy and localization quality of the generated pseudo labels. It effectively segments background points, improving the estimation of pseudo labels’ location, dimension, and orientation. Few training samples always lead to local optima. Introducing multi-modal data in the detector network could enhance the constraints of objects’ features, but 2D images and 3D point clouds have a resolution gap. A motivation for dealing with the resolution gap is that neighboring regions with similar colors and textures in 2D images may exhibit spatial proximity in 3D space. Therefore, a multi-modal network driven by superpixel segmentation is introduced. This network enables effective discrimination between objects in 2D images and 3D point clouds, bridging the resolution gap and leveraging complementary features from both modalities. Experimental results on the KITTI dataset demonstrate the effectiveness of the proposed methods in addressing the challenges associated with weakly-supervised 3D object detection, particularly for categories with ambiguous rectangle boundaries and few instances.

Brain tumor detection based on a novel and high-quality prediction of the tumor pixel distributions

The work presented in this paper is in the area of brain tumor detection. We propose a fast detection system with 3D MRI scans of Flair modality. It performs 2 functions, predicting the gray level distribution and location distribution of the pixels in the tumor regions and generating tumor masks with pixel-wise precision. To facilitate 3D data analysis and processing, we introduce a 2D histogram presentation encompassing the gray-level distribution and pixel-location distribution of a 3D object. In the proposed system, specific 2D histograms highlighting tumor-related features are established by exploiting the left-right asymmetry of a brain structure. A modulation function, generated from the input data of each patient case, is applied to the 2D histograms to transform them into coarsely or finely predicted distributions of tumor pixels. The prediction result helps to identify/remove tumor-free slices. The prediction and removal operations are performed to the axial, coronal and sagittal slice series of a brain image, transforming it into a 3D minimum bounding box of its tumor region. The bounding box is utilized to finalize the prediction and generate a 3D tumor mask. The proposed system has been tested extensively with the data of more than 1200 patient cases in BraTS2018∼2021 datasets. The test results demonstrate that the predicted 2D histograms resemble closely the true ones. The system delivers also very good tumor detection results, comparable to those of state-of-the-art CNN systems with mono-modality inputs. They are reproducible and obtained at an extremely low computation cost and without need for training.