phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | real-time object detection, whose published progress is stated almost entirely in points of AP— “Among the different object detection algorithms, the YOLO (You Only Look Once) framework has stood out for its remarkable balance of speed and accuracy, enabling the rapid and reliable identification of objects in images.” | |
| claim-type | method— “The AP metric is based on precision–recall metrics, handling multiple object categories, and defining a positive prediction using Intersection over Union (IoU).” | |
| method — evidence | — “The AP metric is based on precision–recall metrics, handling multiple object categories, and defining a positive prediction using Intersection over Union (IoU).” | |
| object of study — evidence | — “We present a comprehensive analysis of YOLO’s evolution, examining the innovations and contributions in each iteration from the original YOLO up to YOLOv8, YOLO-NAS, and YOLO with transformers.” | |
| research question — evidence | — “We present a comprehensive analysis of YOLO’s evolution, examining the innovations and contributions in each iteration from the original YOLO up to YOLOv8, YOLO-NAS, and YOLO with transformers.” | |
| contribution — evidence | — “The AP is computed differently in the VOC and in the COCO datasets.” | |
| mechanism | AP is built on the precision-recall curve, and precision and recall trade against each other - detecting more objects raises recall and produces more false positives - so the single number is an area under a curve, not a count— “There is often a tradeoff between precision and recall; for example, increasing the number of detected objects (higher recall) can result in more false positives (lower precision).” | mechanisms[0] |
| mechanism | what counts as a correct detection at all is set by an IoU threshold, and COCO averages over a RANGE of thresholds from 0.5 to 0.95 - so the definition of `correct` is itself a parameter being averaged over— “It measures the overlap between the ground truth and predicted bounding boxes. The COCO benchmark considers multiple IoU thresholds to evaluate the model’s performance at different levels of localization accuracy.” | mechanisms[1] |
| critiques | cross-dataset comparison of object detection performance, since AP means a different computation on VOC and on COCO— “The differences in AP calculation make it hard to directly compare the performance of object detection models across the two datasets.” | claim_relations[0] |
| supports | the COCO AP as the current standard, on the ground that it evaluates localisation at a range of IoU thresholds rather than one— “This dataset includes 80 object categories and uses a more complex method for calculating AP. Instead of using an 11-point interpolation, it uses a 101-point interpolation, i.e., it computes the precision for 101 recall thresholds from 0 to 1 in increments of 0.01. Also, the AP is obtained by averaging over multiple IoU values instead of just… | claim_relations[1] |
| sense | AP / mAP - THE SAME QUANTITY UNDER TWO NAMES, and the review says so in its own first metrics sentence. `Traditionally called mean average precision`; COCO makes no distinction; the paper picks `AP` by fiat. A FIELD'S CENTRAL METRIC WITH AN UNSETTLED NAME— “The average precision (AP), traditionally called mean average precision (mAP), is the commonly used metric for evaluating the performance of object detection models. It measures the average precision across all categories, providing a single value to compare different models. The COCO dataset makes no distinction between AP and mAP. In the… | senses[0] |
| sense | AP, VOC SENSE - 11-point interpolation of the precision-recall curve, averaged over 20 categories, at a single IoU threshold— “This dataset includes 20 object categories. To compute the AP in VOC, we follow the next steps: - For each category, calculate the precision–recall curve by varying the confidence threshold of the model’s predictions. - Calculate each category’s average precision (AP) using an interpolated 11-point sampling of the precision–recall curve. -… | senses[1] |
| sense | AP, COCO SENSE - 101-point interpolation, 80 categories, averaged over IoU thresholds from 0.5 to 0.95 in steps of 0.05. THE SAME SYMBOL AS THE VOC SENSE AND A DIFFERENT QUANTITY, which is why numbers from the two eras of YOLO cannot be lined up— “This dataset includes 80 object categories and uses a more complex method for calculating AP. Instead of using an 11-point interpolation, it uses a 101-point interpolation, i.e., it computes the precision for 101 recall thresholds from 0 to 1 in increments of 0.01. Also, the AP is obtained by averaging over multiple IoU values instead of just… | senses[2] |
| sense | AP50 - the one COCO figure that is a single-threshold measurement, named separately BECAUSE it is the exception. The existence of a special name for the simple case is the clearest sign that the unqualified word is not simple— “This dataset includes 80 object categories and uses a more complex method for calculating AP. Instead of using an 11-point interpolation, it uses a 101-point interpolation, i.e., it computes the precision for 101 recall thresholds from 0 to 1 in increments of 0.01. Also, the AP is obtained by averaging over multiple IoU values instead of just… | senses[3] |
| sense | Non-maximum suppression - post-processes multiple overlapping predictions by retaining the most accurate bounding boxes— “Non-maximum suppression (NMS) is a post-processing technique used in object detection algorithms to reduce the number of overlapping bounding boxes and improve the overall detection quality. Object detection algorithms typically generate multiple bounding boxes around the same object with different confidence scores. NMS filters out redundant… | senses[4] |
| sense | Post-processing technique - NMS is explicitly placed in this method class— “Non-maximum suppression (NMS) is a post-processing technique used in object detection algorithms to reduce the number of overlapping bounding boxes and improve the overall detection quality. Object detection algorithms typically generate multiple bounding boxes around the same object with different confidence scores. NMS filters out redundant… | senses[5] |
| sense | YOLO - completes object detection through a single network pass rather than proposal generation followed by classification— “The name YOLO stands for “You Only Look Once”, referring to the fact that it was able to accomplish the detection task with a single pass of the network, as opposed to previous approaches that either used sliding windows followed by a classifier that needed to run hundreds or thousands of times per image or the more advanced methods that… | senses[6] |
| sense | You Only Look Once - the full phrase encoded by the YOLO acronym— “The name YOLO stands for “You Only Look Once”, referring to the fact that it was able to accomplish the detection task with a single pass of the network, as opposed to previous approaches that either used sliding windows followed by a classifier that needed to run hundreds or thousands of times per image or the more advanced methods that… | senses[7] |
| sense | Single pass - the design choice named by YOLO and contrasted with proposal/classifier pipelines— “The name YOLO stands for “You Only Look Once”, referring to the fact that it was able to accomplish the detection task with a single pass of the network, as opposed to previous approaches that either used sliding windows followed by a classifier that needed to run hundreds or thousands of times per image or the more advanced methods that… | senses[8] |
| sense | Two-steps - first detect candidate object regions, then classify the region proposals— “The name YOLO stands for “You Only Look Once”, referring to the fact that it was able to accomplish the detection task with a single pass of the network, as opposed to previous approaches that either used sliding windows followed by a classifier that needed to run hundreds or thousands of times per image or the more advanced methods that… | senses[9] |
| sense | Anchor boxes - predefined shapes that match prototypical object shapes and receive coordinate/class predictions— “Use anchor boxes to predict bounding boxes. They use a set of prior boxes or anchor boxes, which are boxes with predefined shapes used to match prototypical shapes of objects as shown in Figure 7. Multiple anchor boxes are defined for each grid cell, and the system predicts the coordinates and the class for every anchor box. The size of the… | senses[10] |
| sense | Prior boxes - the review's co-occurring name for anchor boxes— “Use anchor boxes to predict bounding boxes. They use a set of prior boxes or anchor boxes, which are boxes with predefined shapes used to match prototypical shapes of objects as shown in Figure 7. Multiple anchor boxes are defined for each grid cell, and the system predicts the coordinates and the class for every anchor box. The size of the… | senses[11] |
| sense | Dimension clusters - k-means-derived box priors balancing recall against model complexity— “Dimension clusters. Picking good prior boxes helps the network learn to predict more accurate bounding boxes. The authors ran k-means clustering on the training bounding boxes to find good priors. They selected five prior boxes, providing a good tradeoff between recall and model complexity” | senses[12] |
| sense | Objectness score - YOLOv3 logistic score set to one for the anchor with highest ground-truth overlap— “YOLOv3 predicts an objectness score for each bounding box using logistic regression. This score is 1 for the anchor box with the highest overlap with the ground truth and 0 for the rest of the anchor boxes” | senses[13] |
| sense | Backbone architecture - YOLOv2's feature extractor, called Darknet-19— “The backbone architecture used by YOLOv2 is called Darknet-19, containing 19 convolutional layers and 5 max-pooling layers” | senses[14] |
| sense | Darknet-19 - a backbone with nineteen convolutional and five max-pooling layers— “The backbone architecture used by YOLOv2 is called Darknet-19, containing 19 convolutional layers and 5… | senses[15] |
| control | the review derives AP separately for each dataset convention before reporting any model's score, rather than treating the metric as given — held — rules out a reader lining up a YOLOv2 number against a YOLOv5 number - the first is VOC AP and the second is COCO AP, and they are different quantities | controls[0] |
| control | the naming collision is settled explicitly - 'in the rest of this paper, we will refer to this metric as AP' - instead of the two words being used interchangeably — held Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. — rules out a review whose own text is ambiguous about which of two names it means, in a document whose whole purpose is to compare numbers Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[1] |
| control | whether the reported per-version gains are comparable to one another — absent-not-in-staged-text Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. — rules out nothing - the gains are quoted in AP points and the review has just shown AP is two computations, so a 2.3-point gain and a 2.4-point gain are only comparable if both were measured under the same convention, which the review does not state per gain Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[2] |
| weight | cross-dataset comparison of object detection performance, since AP means a different computation on VOC and on COCO — 0.9 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | claim_relations[0] |
| weight | the COCO AP as the current standard, on the ground that it evaluates localisation at a range of IoU thresholds rather than one — 0.8 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | claim_relations[1] |
| note | Non-maximum suppression —is-a→ post-processing technique — The definition explicitly states that NMS is a post-processing technique. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[0] |
| note | single pass —contrasted-with→ two-steps — The review contrasts the single network pass with earlier two-step proposal/classifier methods using 'as opposed to'. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[1] |
| note | backbone architecture —same-as→ Darknet-19 — The review says the YOLOv2 backbone architecture is called Darknet-19. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[2] |
| outcome | control — held Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[0] |
| outcome | control — held Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[1] |
| outcome | control — absent-not-in-staged-text Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[2] |
| no type for this | YOLO → You Only Look Once — YOLO is the acronym that stands for You Only Look Once. Abbreviation expansion is not is-a, part-of, same-as, distinct-from, contrasted-with, operationalised-as, causes, or renamed-across-fields. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[0] |
| limitation | TWO NAMES FOR ONE THING: `average precision (AP), traditionally called mean average precision (mAP)`, and `The COCO dataset makes no distinction between AP and mAP`. The review has to declare which word it will use before it can use it Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[0] |
| limitation | AND ONE NAME FOR TWO COMPUTATIONS: `The AP is computed differently in the VOC and in the COCO datasets` - 11-point interpolation over 20 categories at one IoU threshold, against 101-point interpolation over 80 categories averaged over ten IoU thresholds Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[1] |
| limitation | THE CONSEQUENCE IS STATED PLAINLY AND IT IS ABOUT COMPARABILITY: `The differences in AP calculation make it hard to directly compare the performance of object detection models across the two datasets` Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[2] |
| limitation | the improvements the field reports are quoted in that same word - 2.3 points from simOTA, 2.4 from stronger augmentation - so the increments inherit whatever the number means on the day Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[3] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[0] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[1] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[2] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[3] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[4] |
| quantity | recall thresholds in the COCO interpolation, against 11 in VOC - and 80 categories against 20 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.quantities[0] |
| value source note | quantity — THE DESCRIPTION DRAWS ON THE PAPER BEYOND THIS QUOTE: the VOC's 20 categories are stated just before ('VOC Dataset This dataset includes 20 object categories.'). | content_tags.quantities[0] |
| quantity | the single IoU threshold behind AP50, the one COCO figure that is not averaged over a range Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.quantities[1] |
| variable | confidence threshold Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.variables[0] |
| definition | average precision Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.definitions[0] |
| definition | Intersection over Union Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.definitions[1] |
| concept | average precision — not defined in this paper's own words; the field's ranking metric, computed one way on VOC and another on COCO Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[0] |
| concept | Intersection over Union — not defined in this paper's own words; the overlap between predicted and ground-truth boxes, and the threshold that decides what counts as a correct detection Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[1] |
| concept | precision — not defined in this paper's own words; the accuracy of positive predictions, traded against recall Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[2] |
| concept | recall — not defined in this paper's own words; the proportion of actual positives found, traded against precision Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[3] |
| concept | Non-maximum suppression — not defined in this paper's own words; post-processing that filters overlapping detections to retain the most accurate boxes Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[4] |
| concept | post-processing technique — not defined in this paper's own words; the class of procedure to which the review assigns NMS Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[5] |
| concept | YOLO — not defined in this paper's own words; a detector that completes detection in one network pass Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[6] |
| concept | You Only Look Once — not defined in this paper's own words; the phrase for which the YOLO acronym stands Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[7] |
| concept | single pass — not defined in this paper's own words; YOLO's end-to-end network traversal for detection Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[8] |
| concept | two-steps — not defined in this paper's own words; a region-proposal step followed by classification of those proposals Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[9] |
| concept | anchor boxes — not defined in this paper's own words; predefined box shapes matched to prototypical object shapes at each grid cell Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[10] |
| concept | prior boxes — not defined in this paper's own words; the review's alternative wording for predefined anchor-box shapes Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[11] |
| concept | Dimension clusters — not defined in this paper's own words; k-means clustering of training boxes to choose useful bounding-box priors Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[12] |
| concept | objectness score — not defined in this paper's own words; a logistic score marking the anchor with highest ground-truth overlap Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[13] |
| concept | backbone architecture — not defined in this paper's own words; the YOLOv2 feature-extraction architecture named Darknet-19 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[14] |
| concept | Darknet-19 — not defined in this paper's own words; YOLOv2's 19-convolution, 5-max-pooling backbone Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[15] |
| average precision | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[0] |
| precision | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[1] |
| recall | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[2] |
When we have changed our mind. Some entries in the table begin with a word in capitals and the reason for it. EVIDENCE CORRECTED — the finding still stands, but we had quoted the wrong sentence from the paper, and the quotation has been replaced with the right one. RETRACTED — the entry itself was wrong; it stays on the page, marked, with the reason, rather than being deleted. WITHDRAWN AND REPLACED — the entry was wrong, and a corrected version of it appears in the entries below. Each mark carries its own reason, in brackets, because the reasons differ from one entry to the next. An entry we have WITHDRAWN is removed from this page and kept in the record, with its reason. A line reading Audit and a date is a later re-reading of the entry against its whole sentence: DOUBTFUL means it was kept, and the reason says what is loose about it. Quotation marks inside the tagger's sentence. Where a tag is the tagger's own sentence, a phrase in quotation marks inside it is borrowed from the paper — the tagger's wording around it is not. The paper's own sentence is the evidence quote for that row; on this page it is withheld with the article, and the article is at its publisher. Nothing inside quotation marks anywhere on this page is ours: if it is quoted, it is the paper's. Whose words a concept carries. A concept entry records a term the paper itself uses, so the term, the sense given for it, and its evidence are the paper's own words, quoted where the paper fixes them. This is the opposite of a sense entry in the tag table, which is the tagger's sentence about how the paper uses a term. Where a concept entry could not be grounded in the paper's words, the entry says so on its own row.
Tag layer © the Meta-Tagging Project, CC BY 4.0. The paper text is © its authors/publisher
and is linked, not reproduced. Excerpts shown are short quotations for identification and verification.
· back to index
· concepts · authors
· עברית
© 2026 Shir Sivroni · ORCID 0009-0000-2597-2824