A compact convolutional network for 46-class DHCD Devanagari recognition and reached 99.73%, the highest reported at 15.6x smaller than prior state-of-the-art, effectively reaching the saturation point.
Abstract
We built a compact convolutional network (1.11 M parameters) for 46-class DHCD Devanagari recognition and reached 99.73%, the highest reported at 15.6x smaller than prior state-of-the-art. We have effectively reached the saturation point: every model tested, large teacher ensembles included, hits the same 11-error intrinsic floor. No configuration achieves a statistically clear win under exact McNemar tests with Wilson confidence intervals. Even without knowledge distillation, our student matches the nearest large-model baseline (17.32 M parameters; McNemar $p = 0.345$). Outside of DHCD, zero-shot on CMATERdb digits gives 76.6% and fine-tuning reaches 97.8%; corruption robustness is also far better than large baselines (mean corruption accuracy 75.7% vs. 38.7%). All artifacts are at https://github.com/Ampixa/barnamala.
This paper introduces NepScript Genesis, a Neural Architecture Search (NAS) framework for automated Generative Adversarial Network (GAN) discovery, applied to conditional Devanagari handwritten digit synthesis. We compare five NAS strategies against a carefully constructed Deep Convolutional GAN (DCGAN) baseline (FID=332.28). Architecture selection utilizes a two-stage pipeline guided by a novel domain-aware evaluation metric (Enhanced Score). Results demonstrate that Adaptive Exploration achieves the optimal quality-efficiency trade-off, attaining an FID of 79.12 -- a 76.19% improvement over the baseline -- and the highest mode coverage among the NAS strategies (Recall=0.531) in under one GPU-hour. Furthermore, we demonstrate that incorporating script-specific structural heuristics into the search phase prevents early-stage mode collapse. In a downstream low-resource evaluation, augmenting 250 real training samples per class with GAN-generated digits from the best NAS model improves CNN classification accuracy from 91.0% to 96.5% (+5.5 percentage points), demonstrating that NAS-optimized synthesis produces digits of sufficient quality to benefit practical recognition pipelines when real data is scarce.
Mausam Gurung, Prabin Neupane, S. Acharya· 0 citations
While the top model on OmniDocBench now reaches 96.34% overall on printed-document parsing, the ability of current models to handle challenging handwritten documents remains largely uncharacterized. Existing benchmarks focus on isolated text or formulas, overlook handwritten tables and real-world degradation, and report aggregate accuracy without explaining why models fail. We present WildHandBench, a benchmark containing 500 handwritten documents across three structures (free text, tables, formulas), four languages, and nine real-world scenarios. We introduce a Prior-Driven Error (PDE) metric that quantifies whether errors originate from language priors rather than visual evidence. Evaluating 18 state-of-the-art models together with calibrated human baselines, we find: (1) the best model achieves only 71.85% overall; (2) humans outperform all models yet the gap is narrow (77.09% vs. 71.85%); and (3) model errors are qualitatively different from human errors -- 63-91% of model errors are prior-driven versus only 49% for humans, exposing systematic reliance on language priors that conventional accuracy metrics cannot capture.
Jun Zhang, Qiao Zhao, Cheng Cui et al.· 0 citations
American sign language (ASL) is widely used for communication among the deaf and hard-of-hearing communities. This study aims to develop an optimized deep learning (DL) model for recognizing 36 static ASL gestures representing English alphabets (A–Z) and digits (0–9). A publicly available ASL dataset was used, and five convolutional neural network (CNN) architectures—AlexNet, GoogleNet, Inception V3, MobileNet, and ResNet50—were implemented as baselines. MobileNet was further optimized using the sailfish optimization (SFO) algorithm to fine-tune key hyperparameters and architectural settings. The models were evaluated using accuracy, macro precision, recall, F1-score, specificity, Cohen’s Kappa, Matthews correlation coefficient (MCC), balanced accuracy, Jaccard index, and error rate. The SFO-enhanced MobileNet achieved the highest performance, with 98.28% accuracy, 98.27% macro F1-score, 99.95% specificity, and a 1.72% error rate, outperforming all baselines across metrics. These results demonstrate that SFO optimization significantly improves MobileNet’s ability to classify ASL gestures accurately and efficiently. The proposed model’s high accuracy, robustness, and low inference time (22 ms) make it suitable for real-time sign language interpretation tools and assistive communication devices, supporting broader accessibility and inclusivity.
Sabura Banu Urundai Meeran, S. Kavitha, Balakrishnan Chinthamani et al.· Bulletin of Electrical Engin...· 0 citations
The authors introduce the Edge Suitability Score (ESS), a composite metric that combines normalized accuracy, model size, and inference speed into a single value, weighted at 0.40, 0.35, and 0.25 to reflect their relative importance for microcontroller deployment.
Deeper modern networks outperform the older AlexNet by a wide margin on CIFAR-10, and even a relatively compact ResNet can nearly match the accuracy of a much larger VGG16 in far less time.
Scene text recognition is reported as 89--97% accurate on the six standard benchmarks, and the problem is widely treated as saturated. We present an alternative reading. When the same test images are stratified jointly by ground-truth word rarity and character n-gram novelty against a reference corpus, accuracy at the rare-word x rare-trigram corner of the resulting 5x5 grid drops 10--18 pt below the q3/q3 centre across nine English specialised recognisers, and the same direction (corner below centre) holds on all 13 of 13 (language, model) pairs we test across four writing systems (Latin, Han, Han+kana, Arabic). The drop is not a capacity bottleneck. A 6x vision-backbone scale-up (CLIP4STR-Base 158M ->CLIP4STR-Huge 1.0B, OpenCLIP ViT-H/14 LAION-2B) leads every benchmark in aggregate accuracy yet leaves the stress corner unchanged (86.9 ->86.5, within paired-bootstrap noise). Four converging probes--layer-wise probing, confidence-when-wrong, attention re-balancing, and a cross-script commit-vs-abstain error split--localise the failure to the autoregressive decoder's lexical prior. We then ask how much of the gap existing techniques recover. Of 16 non-architectural mitigations, the largest mean q5/q5 gain is +1.3 pt and none clears the paired-bootstrap noise floor; the only intervention that does is the architectural shift from autoregressive to CTC decoding (SVTRv2, +2.5 pt, p=0.02, n=474). A confidence-routed AR-CTC ensemble adds a directionally consistent +0.6 pt that stays within noise, and its dominant learned coefficient is each model's own minimum-softmax confidence--independently echoing the mechanism above. No configuration we test improves both the compositional corner and aggregate accuracy. The rare-input long tail thus points to architectural change rather than added capacity.
Gen-Pei Zhang· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
