Pretrained vision-language models, such as CLIP, show promising zero-shot performance across a wide variety of datasets. For closed-set classification tasks, however, there is an inherent limitation: CLIP image encoders are typically designed to extract generic image-level features that summarize superfluous or confounding information for the target tasks. This results in degradation of classification performance, especially when objects of interest cover small areas of input images. In this work, we propose CLIP with Guided Cropping (GC-CLIP), where we use an off-the-shelf zero-shot object detection model in a preprocessing step to increase focus of zero-shot classifier to the object of interest and minimize influence of extraneous image regions. We empirically show that our approach improves zero-shot classification results across architectures and datasets, favorably for small objects.
Vision-language models such as CLIP have boosted the performance of open-vocabulary object detection, where the detector is trained on base categories but required to detect novel categories. Existing methods leverage CLIP's strong zero-shot recognition ability to align object-level embeddings with textual embeddings of categories. However, we observe that using CLIP for object-level alignment results in overfitting to base categories, i.e., novel categories most similar to base categories have particularly poor performance as they are recognized as similar base categories. In this paper, we first identify that the loss of critical fine-grained local image semantics hinders existing methods from attaining strong base-to-novel generalization. Then, we propose Early Dense Alignment (EDA) to bridge the gap between generalizable local semantics and object-level prediction. In EDA, we use object-level supervision to learn the dense-level rather than object-level alignment to maintain the local fine-grained semantics. Extensive experiments demonstrate our superior performance to competing approaches under the same strict setting and without using external training resources, i.e., improving the +8.4% novel box AP50 on COCO and +3.9% rare mask AP on LVIS.
This paper aims to remove specular highlights from a single object-level image. Although previous methods have made some progresses, their performance remains somewhat limited, particularly for real images with complex specular highlights. To this end, we propose a three-stage network to address them. Specifically, given an input image, we first decompose it into the albedo, shading, and specular residue components to estimate a coarse specular-free image. Then, we further refine the coarse result to alleviate its visual artifacts such as color distortion. Finally, we adjust the tone of the refined result to match that of the input as closely as possible. In addition, to facilitate network training and quantitative evaluation, we present a large-scale synthetic dataset of object-level images, covering diverse objects and illumination conditions. Extensive experiments illustrate that our network is able to generalize well to unseen real object-level images, and even produce good results for scene-level images with multiple background objects and complex lighting.
We propose a new paradigm to automatically generate training data with accurate labels at scale using the text-to-image synthesis frameworks (e.g., DALL-E, Stable Diffusion, etc.). The proposed approach1 decouples training data generation into foreground object generation, and contextually coherent background generation. To generate foreground objects, we employ a straightforward textual template, incorporating the object class name as input prompts. This is fed into a text-to-image synthesis framework, producing various foreground images set against isolated backgrounds. A foreground-background segmentation algorithm is then used to generate foreground object masks. To generate context images, we begin by creating language descriptions of the context. This is achieved by applying an image captioning method to a small set of images representing the desired context. These textual descriptions are then transformed into a diverse array of context images via a text-to-image synthesis framework. Subsequently, we composite these with the foreground object masks produced in the initial step, utilizing a cut-and-paste method, to formulate the training data. We demonstrate the advantages of our approach on five object detection and segmentation datasets, including Pascal VOC and COCO. We found that detectors trained solely on synthetic data produced by our method achieve performance comparable to those trained on real data (Fig. 1). Moreover, a combination of real and synthetic data yields even much better results. Further analysis indicates that the synthetic data distribution complements the real data distribution effectively. Additionally, we emphasize the compositional nature of our data generation approach in out-of-distribution and zero-shot data generation scenarios. We open-source our code at https://github.com/gyhandy/Text2Image-for-Detection
The introduction of robust backbones, such as Vision Transformers, has improved the performance of object tracking algorithms in recent years. However, these state-of-the-art trackers are computationally expensive since they have a large number of model parameters and rely on specialized hardware (e.g., GPU) for faster inference. On the other hand, recent lightweight trackers are fast but are less accurate, especially on large-scale datasets. We propose a lightweight, accurate, and fast tracking algorithm using Mobile Vision Transformers (MobileViT) as the backbone for the first time. We also present a novel approach of fusing the template and search region representations in the MobileViT backbone, thereby generating superior feature encoding for target localization. The experimental results show that our MobileViT-based Tracker, MVT, surpasses the performance of recent lightweight trackers on the large-scale datasets GOT10k and TrackingNet, and with a high inference speed. In addition, our method outperforms the popular DiMP-50 tracker despite having 4.7 times fewer model parameters and running at 2.8 times its speed on a GPU. The tracker code and models are available at https://github.com/goutamyg/MVT
We present SHIFT3D, a differentiable pipeline for generating 3D shapes that are structurally plausible yet challenging to 3D object detectors. In safety-critical applications like autonomous driving, discovering such novel challenging objects can offer insight into unknown vulnerabilities of 3D detectors. By representing objects with a signed distanced function (SDF), we show that gradient error signals allow us to smoothly deform the shape or pose of a 3D object in order to confuse a downstream 3D detector. Importantly, the objects generated by SHIFT3D physically differ from the baseline object yet retain a semantically recognizable shape. Our approach provides interpretable failure modes for modern 3D object detectors, and can aid in preemptive discovery of potential safety risks within 3D perception systems before these risks become critical failures.
Deep learning approaches to natural language processing have made great strides in recent years. While these models produce symbols that convey vast amounts of diverse knowledge, it is unclear how such symbols are grounded in data from the world. In this paper, we explore the development of a private language for visual data representation by training emergent language (EL) encoders/decoders in both i) a traditional referential game environment and ii) a contrastive learning environment utilizing a within-class matching training paradigm. An additional classification layer utilizing neural machine translation and random forest classification was used to transform symbolic representations (sequences of integer symbols) to class labels. These methods were applied in two experiments focusing on object recognition and action recognition. For object recognition, a set of sketches produced by human participants from real imagery was used (Sketchy dataset) and for action recognition, 2D trajectories were generated from 3D motion capture systems (MOVI dataset). In order to interpret the symbols produced for data in each experiment, gradient-weighted class activation mapping (Grad-CAM) methods were used to identify pixel regions indicating semantic features which contribute evidence towards symbols in learned languages. Additionally, a t-distributed stochastic neighbor embedding (t-SNE) method was used to investigate embeddings learned by CNN feature extractors.
3D object detection using LiDAR point clouds is a fundamental task in the fields of computer vision, robotics, and autonomous driving. However, existing 3D detectors heavily rely on annotated datasets, which are both time-consuming and prone to errors during the process of labeling 3D bounding boxes. In this paper, we propose a Scene Completion Pre-training (SCP) method to enhance the performance of 3D object detectors with less labeled data. SCP offers three key advantages: (1) Improved initialization of the point cloud model. By completing the scene point clouds, SCP effectively captures the spatial and semantic relationships among objects within urban environments. (2) Elimination of the need for additional datasets. SCP serves as a valuable auxiliary network that does not impose any additional efforts or data requirements on the 3D detectors. (3) Reduction of the amount of labeled data for detection. With the help of SCP, the existing state-of-the-art 3D detectors can achieve comparable performance while only relying on 20% labeled data.
Object detection is the task of detecting objects in an image. In this task, the detection of small objects is particularly difficult. Other than the small size, it is also accompanied by difficulties due to blur, occlusion, and so on. Current small object detection methods are tailored to small and dense situations, such as pedestrians in a crowd or far objects in remote sensing scenarios. However, when the target object is small and sparse, there is a lack of objects available for training, making it more difficult to learn effective features. In this paper, we propose a specialized method for detecting a specific category of small objects; birds. Particularly, we improve the features learned by the neck; the sub-network between the backbone and the prediction head, to learn more effective features with a hierarchical design. We employ Swin Transformer to upsample the image features. Moreover, we change the shifted window size for adapting to small objects. Experiments show that the proposed Swin Transformer-based neck combined with CenterNet can lead to good performance by changing the window sizes. We further find that smaller window sizes (default 2) benefit mAPs for small object detection.
Detecting small objects is often impeded by blurriness and low resolution, which poses substantial challenges for accurately detecting and localizing such objects. In addition, conventional feature extraction methods usually face difficulties in capturing effective representations for these entities, as down-sampling and convolutional operations contribute to the blurring of small object details. To tackle these challenges, this study introduces an approach for detecting tiny objects through ensemble fusion, which leverages the advantages of multiple diverse model variants and combines their predictions. Experimental results reveal that the proposed method effectively harnesses the strengths of each model via ensemble fusion, leading to enhanced accuracy and robustness in small object detection. Our model achieves the highest score of 0.776 in terms of average precision (AP) at an IoU threshold of 0.5 in the MVA Challenge on Small Object Detection for Birds.
In this paper, we propose and realize a new deep learning architecture for discovering symbolic representations for objects and their relations based on the self-supervised continuous interaction of a manipulator robot with multiple objects on a tabletop environment. The key feature of the model is that it can handle a changing number number of objects naturally and map the object-object relations into symbolic domain explicitly. In the model, we employ a self-attention layer that computes discrete attention weights from object features, which are treated as relational symbols between objects. These relational symbols are then used to aggregate the learned object symbols and predict the effects of executed actions on each object. The result is a pipeline that allows the formation of object symbols and relational symbols from a dataset of object features, actions, and effects in an end-to-end manner. We compare the performance of our proposed architecture with state-of-the-art symbol discovery methods in a simulated tabletop environment where the robot needs to discover symbols related to the relative positions of objects to predict the observed effect successfully. Our experiments show that the proposed architecture performs better than other baselines in effect prediction while forming not only object symbols but also relational symbols. Furthermore, we analyze the learned symbols and relational patterns between objects to learn about how the model interprets the environment. Our analysis shows that the learned symbols relate to the relative positions of objects, object types, and their horizontal alignment on the table, which reflect the regularities in the environment.
We propose a novel framework for interactive class-agnostic object counting, where a human user can interactively provide feedback to improve the accuracy of a counter. Our framework consists of two main components: a user-friendly visualizer to gather feedback and an efficient mechanism to incorporate it. In each iteration, we produce a density map to show the current prediction result, and we segment it into non-overlapping regions with an easily verifiable number of objects. The user can provide feedback by selecting a region with obvious counting errors and specifying the range for the estimated number of objects within it. To improve the counting result, we develop a novel adaptation loss to force the visual counter to output the predicted count within the user-specified range. For effective and efficient adaptation, we propose a refinement module that can be used with any density-based visual counter, and only the parameters in the refinement module will be updated during adaptation. Our experiments on two challenging class-agnostic object counting benchmarks, FSCD-LVIS and FSC-147, show that our method can reduce the mean absolute error of multiple state-of-the-art visual counters by roughly 30% to 40% with minimal user input. Our project can be found at https://yifehuang97.github.io/ICACountProjectPage/.
Automated detection of Gallbladder Cancer (GBC) from Ultrasound (US) images is an important problem, which has drawn increased interest from researchers. However, most of these works use difficult-to-acquire information such as bounding box annotations or additional US videos. In this paper, we focus on GBC detection using only image-level labels. Such annotation is usually available based on the diagnostic report of a patient, and do not require additional annotation effort from the physicians. However, our analysis reveals that it is difficult to train a standard image classification model for GBC detection. This is due to the low inter-class variance (a malignant region usually occupies only a small portion of a US image), high intra-class variance (due to the US sensor capturing a 2D slice of a 3D object leading to large viewpoint variations), and low training data availability. We posit that even when we have only the image level label, still formulating the problem as object detection (with bounding box output) helps a deep neural network (DNN) model focus on the relevant region of interest. Since no bounding box annotations is available for training, we pose the problem as weakly supervised object detection (WSOD). Motivated by the recent success of transformer models in object detection, we train one such model, DETR, using multi-instance-learning (MIL) with self-supervised instance selection to suit the WSOD task. Our proposed method demonstrates an improvement of AP and detection sensitivity over the SOTA transformer-based and CNN-based WSOD methods. Project page is at https://gbc-iitd.github.io/wsod-gbc
We introduce the task of localizing a flexible number of objects in real-world 3D scenes using natural language descriptions. Existing 3D visual grounding tasks focus on localizing a unique object given a text description. However, such a strict setting is unnatural as localizing potentially multiple objects is a common need in real-world scenarios and robotic tasks (e.g., visual navigation and object rearrangement). To address this setting we propose Multi3DRefer, generalizing the ScanRefer dataset and task. Our dataset contains 61926 descriptions of 11609 objects, where zero, single or multiple target objects are referenced by each description. We also introduce a new evaluation metric and benchmark methods from prior work to enable further investigation of multi-modal 3D scene understanding. Furthermore, we develop a better baseline leveraging 2D features from CLIP by rendering object proposals online with contrastive learning, which outperforms the state of the art on the ScanRefer benchmark.
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.
Transformers have rapidly gained popularity in computer vision, especially in the field of object recognition and detection. Upon examining the outcomes of state-of-the-art object detection methods, we noticed that transformers consistently outperformed well-established CNN-based detectors in almost every video or image dataset. While transformer-based approaches remain at the forefront of small object detection (SOD) techniques, this paper aims to explore the performance benefits offered by such extensive networks and identify potential reasons for their SOD superiority. Small objects have been identified as one of the most challenging object types in detection frameworks due to their low visibility. We aim to investigate potential strategies that could enhance transformers' performance in SOD. This survey presents a taxonomy of over 60 research studies on developed transformers for the task of SOD, spanning the years 2020 to 2023. These studies encompass a variety of detection applications, including small object detection in generic images, aerial images, medical images, active millimeter images, underwater images, and videos. We also compile and present a list of 12 large-scale datasets suitable for SOD that were overlooked in previous studies and compare the performance of the reviewed studies using popular metrics such as mean Average Precision (mAP), Frames Per Second (FPS), number of parameters, and more. Researchers can keep track of newer studies on our web page, which is available at \url{https://github.com/arekavandi/Transformer-SOD}.
To date, most instance segmentation approaches are based on supervised learning that requires a considerable amount of annotated object contours as training ground truth. Here, we propose a framework that searches for the target object based on a shape prior. The shape prior model is learned with a variational autoencoder that requires only a very limited amount of training data: In our experiments, a few dozens of object shape patches from the target dataset, as well as purely synthetic shapes, were sufficient to achieve results en par with supervised methods with full access to training data on two out of three cell segmentation datasets. Our method with a synthetic shape prior was superior to pre-trained supervised models with access to limited domain-specific training data on all three datasets. Since the learning of prior models requires shape patches, whether real or synthetic data, we call this framework semi-supervised learning.
Designing metrics for evaluating instance segmentation revolves around comprehensively considering object detection and segmentation accuracy. However, other important properties, such as sensitivity, continuity, and equality, are overlooked in the current study. In this paper, we reveal that most existing metrics have a limited resolution of segmentation quality. They are only conditionally sensitive to the change of masks or false predictions. For certain metrics, the score can change drastically in a narrow range which could provide a misleading indication of the quality gap between results. Therefore, we propose a new metric called sortedAP, which strictly decreases with both object- and pixel-level imperfections and has an uninterrupted penalization scale over the entire domain. We provide the evaluation toolkit and experiment code at https://www.github.com/looooongChen/sortedAP.
Underwater object detection faces the problem of underwater image degradation, which affects the performance of the detector. Underwater object detection methods based on noise reduction and image enhancement usually do not provide images preferred by the detector or require additional datasets. In this paper, we propose a plug-and-play Underwater joint image enhancement Module (UnitModule) that provides the input image preferred by the detector. We design an unsupervised learning loss for the joint training of UnitModule with the detector without additional datasets to improve the interaction between UnitModule and the detector. Furthermore, a color cast predictor with the assisting color cast loss and a data augmentation called Underwater Color Random Transfer (UCRT) are designed to improve the performance of UnitModule on underwater images with different color casts. Extensive experiments are conducted on DUO for different object detection models, where UnitModule achieves the highest performance improvement of 2.6 AP for YOLOv5-S and gains the improvement of 3.3 AP on the brand-new test set (URPCtest). And UnitModule significantly improves the performance of all object detection models we test, especially for models with a small number of parameters. In addition, UnitModule with a small number of parameters of 31K has little effect on the inference speed of the original object detection model. Our quantitative and visual analysis also demonstrates the effectiveness of UnitModule in enhancing the input image and improving the perception ability of the detector for object features.
Multiple object tracking (MOT) tends to become more challenging when severe occlusions occur. In this paper, we analyze the limitations of traditional Convolutional Neural Network-based methods and Transformer-based methods in handling occlusions and propose DNMOT, an end-to-end trainable DeNoising Transformer for MOT. To address the challenge of occlusions, we explicitly simulate the scenarios when occlusions occur. Specifically, we augment the trajectory with noises during training and make our model learn the denoising process in an encoder-decoder architecture, so that our model can exhibit strong robustness and perform well under crowded scenes. Additionally, we propose a Cascaded Mask strategy to better coordinate the interaction between different types of queries in the decoder to prevent the mutual suppression between neighboring trajectories under crowded scenes. Notably, the proposed method requires no additional modules like matching strategy and motion state estimation in inference. We conduct extensive experiments on the MOT17, MOT20, and DanceTrack datasets, and the experimental results show that our method outperforms previous state-of-the-art methods by a clear margin.