object detection
PulseAugur coverage of object detection — every cluster mentioning object detection across labs, papers, and developer communities, ranked by signal.
- used by autonomous driving 90%
- instance of Ms Coco 90%
- instance of image classification 70%
- used by PASCAL-VOC 70%
- instance of computer vision 70%
- affiliated with autonomous driving 70%
- used by alphaXiv 70%
- competes with semantic segmentation 60%
- used by DagsHub 60%
- instance of instance segmentation 60%
- competes with image classification 50%
2 day(s) with sentiment data
-
Spiking Patches tokenization boosts event camera efficiency by 10x
Researchers have developed a new method called Spiking Patches for processing data from event cameras, which capture asynchronous and sparse visual information. This novel tokenization technique preserves the unique pro…
-
New framework ReVISE enhances MLLM reliability in visual tasks
Researchers have developed a new framework called ReVISE to improve the reliability of multimodal large language models (MLLMs) when using external tools for visual tasks. This framework enables MLLMs to verify the outp…
-
New method adapts Vision Transformers for faster object detection
Researchers have developed a new method called Detector-Interface Distillation (DiD) to adapt Vision Transformers (ViTs) from Softmax attention to linear attention for object detection tasks. This label-free approach fo…
-
New HCFormer Architecture Uses Hyperbolic Clustering for Interpretable Vision Models
Researchers have developed HCFormer, a new vision backbone architecture that utilizes hyperbolic hierarchical clustering for visual representation learning. This approach, named ClusterMixer, offers a more interpretable…
-
New AdROD system enhances adversarial robustness for autonomous driving object detection
Researchers have developed AdROD, a novel defense system designed to enhance the adversarial robustness of object detection in autonomous vehicles. AdROD utilizes low-rank HyperNetworks, significantly reducing computati…
-
Wireless 5G system enhances human-robot collaboration in industry
Researchers have developed a wireless, 5G-based system to enhance human-robot collaboration (HRC) in industrial settings, particularly for tasks requiring frequent workcell rearrangements. The system utilizes a novel ba…
-
FaithIR framework enhances infrared image super-resolution for machine perception
Researchers have introduced FaithIR, a novel framework designed to improve infrared image super-resolution (IISR) for enhanced machine perception. Unlike previous methods that often introduce artificial textures or dist…
-
Semantic Segmentation: Pixel-Level Understanding in Computer Vision
Semantic segmentation is a computer vision technique that assigns a specific class label to every pixel within an image. This process enables models to create detailed maps, distinguishing elements like roads, people, o…
-
AI systems can now detect objects more cautiously in poor image quality
Researchers have developed a new method to improve the reliability of AI systems, particularly in automated driving, when faced with poor-quality image data. The approach involves a "fail-degraded" system that lowers th…
-
Apple unveils GH-ESD for discovering vision model errors
Apple's Machine Learning Research team has introduced GH-ESD, a novel framework for discovering instance-level error slices in vision tasks. This approach reformulates slice discovery as grounded hypothesis generation a…
-
Multi-Object Tracking: Giving AI Systems Memory Beyond Object Detection
Multi-Object Tracking (MOT) is an advancement beyond object detection, providing identity, memory, and historical context to recognized objects within video streams. This is crucial for applications like autonomous driv…
-
New attack targets event-based SNNs, increasing latency by 38%
Researchers have developed a new availability backdoor attack called Event Burst Trigger (EBT) specifically for event-based Spiking Neural Networks (SNNs) used in object detection. This attack injects triggers into trai…
-
Machine learning evaluation metrics explained: Accuracy, IoU, mAP, and more
Evaluation metrics are essential for assessing machine learning model performance, particularly in object detection tasks. Key metrics include accuracy, which can be misleading on imbalanced datasets, and the confusion …
-
New framework evaluates UAV detection and tracking in synthetic fog
Researchers have developed a new framework for evaluating how well unmanned aerial vehicles (UAVs) can be detected and tracked in foggy conditions. This framework uses synthetic fog generated from real-world images to t…
-
Federated learning enables collaborative drone object detection without data centralization
Researchers have developed a federated learning approach for object detection, enabling drones to collaboratively train a shared model without centralizing their data. This method addresses privacy and bandwidth challen…
-
Synthetic fog pipeline reduces data needs for autonomous vehicle object detection
Researchers have developed Clear2Fog (C2F), a physics-based pipeline designed to generate synthetic fog for training object detection models. This method aims to improve the safety of autonomous vehicles by addressing t…
-
New GoodQ method uses generative models for zero-shot object detector quantization
Researchers have developed GoodQ, a new pipeline for Zero-Shot Quantization-Aware Training (ZSQ-OD) that leverages off-the-shelf generative models to create training datasets. This method addresses challenges such as de…
-
New Key-Correlated Layer Attention offers linear complexity for neural networks
Researchers have developed Key-Correlated Layer Attention (KCLA), a novel mechanism designed to improve how different layers within a neural network interact. KCLA addresses the quadratic computational complexity of tra…
-
TaskTok framework enhances downstream vision tasks via selective token restoration
Researchers have introduced TaskTok, a novel framework designed for Task-Driven Image Restoration (TDIR). Unlike traditional methods that focus on perceptual quality, TDIR aims to improve the performance of subsequent h…
-
PrototypeNAS accelerates DNN design for microcontrollers
Researchers have developed PrototypeNAS, a novel zero-shot neural architecture search method designed to rapidly create efficient deep neural networks (DNNs) for microcontroller units (MCUs). This method automates the s…