Computer Vision

A hybrid CNN-GLCM-SVM fusion framework achieved 95.0% accuracy, 93.8% sensitivity, and 96.1% specificity in automated four-class cataract severity classification from standard consumer-grade eye photographs, outperformin…

A new training-free framework called Surprise Forcing improves long-video generation by dynamically allocating resources to key visual evidence and skipping unnecessary denoising steps. Developed by researchers and publi…

A new survey from arXiv (2607.19011v1) finds that multimodal humor in memes, cartoons, and comics remains difficult for AI systems because intended meaning depends on non-literal mechanisms, shared cultural knowledge, an…

Researchers present AniGS, a method for scene-level animation of 3D Gaussian Splatting reconstructions that adds subtle, distributed dynamics like vegetation motion while preserving rigid structures, using a time-conditi…

ECoNGS, an efficient compressive neural Gaussian splatting framework for volume visualization, improves reconstruction quality by up to 2.2 dB in PSNR while reducing model size by up to 6.1x and training time by up to 5.…

A new benchmark, MissingBench-Verified, reveals that ten leading vision-language models consistently fail to detect missing object parts, with failure rates persisting even when external tool evidence contradicts the mod…

A deep learning ensemble developed from 128,621 CT examinations at nine Japanese institutions estimated adult sex, age, height, and weight from coronal digitally reconstructed radiographs with high accuracy, achieving 0.…

Researchers introduce Fluid-SDF, a differentiable Constructive Solid Geometry framework that models shapes using explicit geometric primitives with under 100 parameters, achieving comparable or superior intersection-over…

A new study from arXiv (2607.18540v1) reveals that post-training quantization (PTQ) on edge robotics platforms creates a Quantization-Induced Robustness Gap, degrading reliability under distribution shifts like sensor no…

Deep learning models for breast MRI tumor classification suffer near-chance external accuracy (0.5048–0.5265) when dataset origin is perfectly correlated with label, but a class-wise dataset mixing method using Duke Brea…

A new AI framework called DeforM, developed by researchers, improves physics-aware video generation by using a VLM-guided physical reasoning module to identify dynamic regions and apply spatial-temporal masks, outperform…

A new study from arXiv (2607.18695v1) finds that descriptors generated by large language models (LLMs) for zero-shot classification carry little visual evidence, collapsing ImageNet accuracy from 59.5% to 15.5% when the …

Researchers propose DuSPiT, a dual-branch sub-patch pixel diffusion transformer that separates global structural reasoning from local appearance modeling in pixel-space image generation, achieving richer details and a be…

Researchers propose IGGT4D, a streaming instance-grounded geometry Transformer for online 4D scene understanding that unifies camera motion, scene geometry, and consistent instance features within a single feed-forward f…

Researchers at Huazhong University of Science and Technology and Alibaba Group's Wan Team have developed FilmWorld, a system that generates films from novels by modeling a dynamic cinematic world as a five-tuple with per…

Hans Moravec's paradox, articulated in 1988, observes that computers easily master adult-level abstract reasoning like chess but struggle with perception and mobility skills that are effortless for humans, such as recogn…

Researchers have developed SEAM-V, a hybrid-decoupled vector execution architecture for RISC-V Vector Extension (RVV) that achieves a geometric-mean speedup of 1.34x across 17 representative kernels compared to an Ara-ba…

A developer's benchmark of Google's Gemini 3.5 Flash Lite for human detection in a home surveillance system found the new model produced one false positive in 30 runs, while the existing Gemini 3.1 Flash Lite had zero fa…

A new study from Carnegie Mellon University finds that AI vision systems like CLIP often generate image captions based on stereotypes associated with the image's label rather than by actually analyzing the visual content…

Using depth maps instead of traditional storyboards improves foreground-background separation in AI video generation by providing explicit spatial context that AI models need to avoid common problems like blurring and un…