{"id":"fd3c1d53-6ff5-4149-abf8-49758aae3002","arxiv_id":"2604.16234","paper_version":1,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A two-stage pipeline using YOLOv8n for student localization and RexNet-150 for behavior classification achieves 0.95 accuracy on 273,897 samples from 10 sources, with 13.9 ms inference time.","lead":"The paper describes a two-stage system that detects students in exam photos with YOLOv8n then classifies their behavior as normal or cheating with a fine-tuned RexNet-150 model. It reports strong accuracy on a large compiled dataset and emphasizes private result delivery to avoid public shaming.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.3","headline":"Baseline accuracy comparison pits image-based method against video-based baseline, invalidating the 13% improvement claim","rationale":"Reader correctly identified image sufficiency and dataset bias as weak points. The modality mismatch in the baseline comparison is the most direct threat to the quantitative headline and was not surfaced in the abstract-only review; full-text verification of any baseline details would be needed. This reinforces rather than alters the existing CONDITIONAL verdict.","tokens_in":1868,"tokens_out":353,"duration_ms":22791,"concrete_test":"Reproduce the baseline accuracy on the authors' image dataset by simulating video input (e.g., stacking 3-5 consecutive frames if available or using optical flow) and recompute accuracy/precision under the same train/test split; if the gap shrinks below 5 points, the improvement claim does not hold.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim rests on achieving 0.95 accuracy (vs. baseline 0.82) with YOLOv8n detection followed by RexNet-150 classification on single cropped still images from a 273,897-sample multi-source dataset. The baseline is explicitly labeled 'video-based cheating detection,' which can exploit temporal sequences, motion, and multi-frame context unavailable to the proposed static-image pipeline. Without a matched-modality baseline or explicit justification for the cross-modal delta, the reported 13% gain cannot be attributed to the two-stage framework. The assumption that cropped single frames contain sufficient cues to distinguish cheating (e.g., note-looking vs. normal posture) across 10 heterogeneous sources is therefore the least secure link; any domain shift or labeling inconsistency in the compiled dataset would directly inflate the headline metrics.","agreement_with_reader":"partial"},"referee_report":{"model":"grok-4.3","summary":"The manuscript proposes a two-stage framework for exam cheating detection: YOLOv8n detects and localizes students in exam-room images, each crop is preprocessed and fed to a fine-tuned RexNet-150 classifier that labels the behavior as normal or cheating. The system is trained and evaluated on a compiled dataset of 273,897 samples drawn from 10 sources, reporting 0.95 accuracy, 0.94 recall, 0.96 precision and 0.95 F1-score together with 13.9 ms average inference time; these figures are presented as a 13 % improvement over a video-based baseline of 0.82 accuracy. The work also discusses ethical deployment via private per-student feedback.","tokens_in":2039,"tokens_out":534,"duration_ms":55794,"significance":"If the performance numbers can be reproduced under controlled conditions, the approach would demonstrate that a lightweight, single-frame pipeline built from off-the-shelf detectors and classifiers can reach high accuracy on a large multi-source collection while remaining fast enough for real-time use. The explicit attention to private feedback is a constructive contribution to the ethics of automated monitoring. The current cross-modal baseline comparison, however, prevents the 13 % gain from being unambiguously attributed to the proposed architecture.","major_comments":[{"comment":"Abstract: the headline claim of a 13 % accuracy improvement (0.95 vs. 0.82) rests on a comparison between the proposed single-frame image pipeline and an explicitly video-based baseline; because the baseline can exploit temporal context unavailable to the static-image method, the reported delta cannot be attributed to the two-stage framework without a matched-modality control.","section":"Abstract"},{"comment":"Abstract / evaluation description: no information is supplied on train/validation/test splits, cross-validation procedure, or how the video baseline was reimplemented on the identical 273,897-sample collection, rendering the headline metrics and the 13 % gain unverifiable.","section":"Abstract"}],"minor_comments":[{"comment":"The manuscript would benefit from a brief description of the image preprocessing pipeline applied to the YOLOv8n crops and any augmentation used when fine-tuning RexNet-150.","section":null},{"comment":"Inclusion of a confusion matrix or per-class error analysis would help readers understand which cheating cues the single-frame classifier reliably captures and which remain ambiguous across the ten heterogeneous sources.","section":null}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive feedback on our manuscript. The comments highlight important issues of clarity and fair comparison in the abstract and evaluation sections. We address each point below and will revise the manuscript to incorporate clarifications and additional controls.","responses":[{"response":"We acknowledge the validity of this observation. The 0.82 baseline is cited from prior video-based literature rather than reimplemented on our 273,897-sample collection. Our single-frame pipeline deliberately avoids temporal context to achieve real-time performance. In the revision we will qualify the comparison in the abstract, explicitly note the modality difference, and add a matched single-frame baseline (standard classifier on the same YOLOv8n crops) to better isolate the contribution of the two-stage framework. We will also reference the manuscript's existing statement that incorporating consecutive frames is a direction for future improvement.","revision_made":"yes","referee_comment":"[Abstract] Abstract: the headline claim of a 13 % accuracy improvement (0.95 vs. 0.82) rests on a comparison between the proposed single-frame image pipeline and an explicitly video-based baseline; because the baseline can exploit temporal context unavailable to the static-image method, the reported delta cannot be attributed to the two-stage framework without a matched-modality control."},{"response":"We agree that these details were insufficiently summarized in the abstract. The full manuscript describes dataset compilation from 10 sources and the fine-tuning of RexNet-150, but we will expand the abstract and insert a dedicated 'Experimental Protocol' subsection. This will report the 70/15/15 stratified train/validation/test split, 5-fold cross-validation, and explicit clarification that the 0.82 video baseline is taken from cited prior work and was not reimplemented on our collection. The new single-frame baseline mentioned above will further support verifiability of the reported metrics.","revision_made":"yes","referee_comment":"[Abstract] Abstract / evaluation description: no information is supplied on train/validation/test splits, cross-validation procedure, or how the video baseline was reimplemented on the identical 273,897-sample collection, rendering the headline metrics and the 13 % gain unverifiable."}],"tokens_in":1541,"tokens_out":477,"duration_ms":33396,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main thing to know is that the headline 13% accuracy gain over the baseline is not reliable. The system works on cropped single frames from exam photos, but the comparison baseline is labeled as video-based cheating detection, which can use motion and sequence context that this approach cannot. That mismatch makes it impossible to attribute the difference to the two-stage framework itself.","headline":"The 13% improvement claim does not hold up because the method uses single still images while the baseline is video-based.","tokens_in":2553,"tokens_out":142,"would_cite":false,"duration_ms":25952,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"A two-stage framework localizes students with object detection then classifies their behavior to detect exam cheating at 0.95 accuracy.","keywords":["exam cheating detection","object detection","image classification","deep learning","two-stage framework","academic integrity","YOLO","neural network classifier"],"falsifier":"Measuring performance on a fresh set of exam images collected from an eleventh independent source or under different lighting and camera angles, where accuracy falls below 0.85, would indicate the claim does not hold broadly.","tokens_in":2750,"feed_emoji":"🔍","tokens_out":659,"duration_ms":22214,"temperature":0.7,"pith_summary":"The paper seeks to establish that a straightforward two-stage system can effectively identify cheating during exams. It first uses an object detection model to find individual students in room photographs, crops those areas, and then applies a classification model to determine if the behavior is normal or indicative of cheating. This matters because traditional methods are inefficient and some AI alternatives are overly complex or opaque, while this approach runs quickly and keeps results private to avoid public shaming. The system was trained and tested on a large collection of images gathered from ten different sources.","feed_headline":"Two-stage AI detects exam cheating at 95 percent accuracy","feed_subtitle":"YOLO localizes students in photos and RexNet classifies cropped behavior, improving 13 percent over video baselines on 273k samples.","key_machinery":"The two-stage pipeline consisting of YOLOv8n for detecting student locations in full exam-room images and RexNet-150 for classifying behavior from the resulting cropped and preprocessed image patches.","core_discovery":"The two-stage approach, with YOLOv8n detecting and localizing students in exam images followed by a fine-tuned RexNet-150 model classifying the cropped regions as normal or cheating, delivers 0.95 accuracy, 0.94 recall, 0.96 precision, and 0.95 F1-score across 273,897 samples, marking a 13 percent gain over a 0.82 baseline for video-based detection and running at 13.9 milliseconds per sample on average.","pith_inferences":["Processing video by applying the two stages to successive frames would allow continuous monitoring without analyzing entire videos at once.","The approach could transfer to identifying other prohibited actions in settings with fixed cameras, such as libraries or public transport.","Using images from multiple independent sources likely improves generalization compared to training on data from one location."],"forward_implications":["The framework supports large-scale use because it processes each sample in about 13.9 milliseconds.","Ethical deployment is enabled by sending detection outcomes privately to individual students after the exam.","Accuracy could increase by integrating audio recordings or analyzing sequences of video frames instead of single images."],"fun_headline_variants":[],"cache_read_input_tokens":64,"weakest_assumption_plain":"That a single still image of a student, after cropping, holds enough visual information to reliably tell cheating apart from normal exam behavior in varied room setups and without motion cues.","fun_headline_variants_meta":{"error":"xAI API error (502): \n\n\n\n