Pith. sign in

REVIEW 4 major objections 5 minor 37 references

Towards Smart Point-and-Shoot Photography

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims a smartphone camera can guide a user live—by suggesting yaw and pitch rotations—to a better-composed shot of a scene, using a 320K-image pose-labeled dataset and a two-task mixture-of-experts model.

desk verdict New task and a useful dataset, but the objective numbers only show that CPAM imitates CCQA; the user study is the only evidence of real photographic improvement, and it is too small to carry that weight. read the letter →

arxiv 2505.03638 v1 pith:PZHAULB7 submitted 2025-05-06 cs.CV

Jiawan Li , Fei Zhou , Zhipeng Zhong , Jiongzhi Lin , Guoping Qiu This is my paper · ORCID
classification cs.CV
keywords smartpoint-and-shootphotographycameraposeadjustmentimagecompositionassessment360-degreestreetviewdatasetCLIP-basedqualitymixture-of-expertsmulti-tasklearningpseudo-labelgenerationliveviewfinderguidance
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes Smart Point-and-Shoot (SPAS), a system that tells a photographer, live at the scene, whether the current composition can be improved and, if so, how much to rotate the camera left or right (yaw) and up or down (pitch) to reach a better view. To train it, the authors build PCARD, a dataset of 320K perspective views with known camera poses extracted from 4,000 spherical street panoramas, and assign each view a pseudo composition score using CCQA, a CLIP-based model with learnable quality-word prompts. A camera pose adjustment model (CPAM), a two-task mixture-of-experts network, then predicts a binary "adjust or not" decision and, when adjustment is needed, the two pose angles. If the system works as claimed, ordinary smartphone users could be guided to compose better photographs without knowing photography rules, and composition guidance would move from post-hoc cropping to live camera control.

What carries the argument

The argument rides on three coupled mechanisms. First, an equirectangular-to-sphere mapping converts each 360-degree panorama into a unit sphere, so a virtual pinhole camera at the center can generate perspective views with exactly known longitude and latitude poses; sampling is constrained by content preservation (overlap with the initial view) and local redundancy (an 8-neighbor Moore neighborhood with 5-degree steps), yielding about 81 candidate views per scene. Second, CCQA assigns each view a pseudo score in the five-level range {bad, poor, fair, good, perfect} using a CLIP image-text encoder with learnable prompt embeddings and softmax cosine-similarity weights; trained on GAICv2 cropping scores, it provides the ranking that generates adjustment labels through an adaptive threshold at the top 25% of candidate scores. Third, CPAM is a multi-gate mixture-of-experts network whose suggestion head is trained with cross-entropy on all samples and whose adjustment head is trained only on samples flagged for adjustment, using cosine-similarity plus magnitude losses so the predicted $(\Delta\theta, \Delta\varphi)$ matches both direction and size. The expert gates let the two tasks share features while separating their gradients, which is meant to resolve the conflict between a full-sample classification task and a subset-only regression task.

What would settle it

Take a held-out set of real street scenes not in PCARD, capture or render views before and after CPAM's suggested yaw and pitch adjustments, and have independent human raters blindly choose the better composition. If the adjusted view is not preferred significantly more often than chance, or is preferred less than the roughly 64% rate reported in the paper, the central claim fails. A second check would compare CCQA's scores on PCARD views with direct human pairwise preferences; low agreement would invalidate the pseudo-label step that generates every training label.

Watch

Extended reading notes

Core claim

The central claim is that a view's composition quality can be improved by changing camera pose, and that this improvement can be predicted from a single image. Concretely, CPAM first determines if the current view can be further improved; if yes, it outputs a yaw adjustment $\Delta\theta$ and a pitch adjustment $\Delta\varphi$ such that photographing after applying the suggestion yields a better-composed shot. The paper supports the claim with objective metrics on the PCARD test split and with a user study in which 64% of raters preferred the adjusted image over the original when a suggestion was given, while 82% approved decisions not to suggest an adjustment.

Load-bearing premise

The whole system rests on CCQA's pseudo-labels being a faithful measure of photographic composition; if the CLIP-based scorer trained on GAICv2 cropping scores does not match what humans prefer in real street-scene views, CPAM is only learning to imitate a flawed labeler.

Editorial extensions

If this is right

  • If CPAM's predictions are correct, a phone viewfinder can display a live instruction such as "turn left 10 degrees, tilt up 5 degrees," turning composition guidance into a real-time camera control rather than a cropping step after capture.
  • The PCARD dataset's known camera poses make it reusable for other pose-conditioned tasks, such as view synthesis, 360-degree navigation, or aesthetic ranking of viewpoints.
  • CCQA's ability to rank near-identical neighboring views suggests the learnable-prompt scoring method can be applied to other subtle-difference quality judgments where categorical text levels are too coarse.
  • Because the evaluation computes IoU on the spherical panorama surface, composition improvement can be measured geometrically across the full viewing sphere, not only on flat cropped images.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension the paper does not develop is using the same pipeline to recommend small camera translations in addition to rotations, which would require depth information but could further improve composition beyond what yaw and pitch changes can achieve.
  • The approach implicitly assumes composition quality is a smooth function over the view sphere; if true, the discrete 5-degree sampling and local-redundancy pruning could be replaced by continuous optimization, and one could test whether smaller steps yield even better final views.
  • Because CCQA is trained on cropping preferences from GAICv2, the pseudo-labels may inherit a bias toward centered, tightly framed subjects; a direct test would be human pairwise comparison on the PCARD street views themselves, without any cropping context.
  • The same pseudo-labeling pipeline could be adapted to recommend where to point a camera in other immersive settings, such as indoor 360-degree walkthroughs or user-generated virtual tours, where composition preferences may differ from street scenes.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper presents SPAS, a system that suggests yaw and pitch adjustments to a smartphone photographer before the shot is taken. It introduces PCARD, a dataset of 320K perspective views rendered from 4,000 Google Street View panoramas, with camera poses and pseudo-labels produced by a CLIP-based composition quality assessor (CCQA) trained on GAICv2. A camera pose adjustment model (CPAM) with a two-task mixture-of-experts architecture predicts whether adjustment is needed and, if so, the required (Δθ, Δφ). The authors report ablations, objective metrics on PCARD, a generalization experiment for CCQA on CPC, and a small user study.

Significance. The problem formulation is original and practically motivated: existing composition work is mostly post-hoc cropping, not live pose guidance. The dataset with explicit camera-pose labels is a potentially useful resource. CCQA's learnable text prompts and CPAM's gated mixture-of-experts are reasonable technical choices. However, the central evaluation is built on CCQA pseudo-labels, and CCQA's validity for full-frame Street View views is not independently established; the current human evidence is too limited to close that gap. I find the stress-test concern that CCQA pseudo-label validity is load-bearing to be accurate. If the labeler issue is addressed with a proper human study, the contributions could be significant; as presented, the paper supports a demonstration that CPAM imitates CCQA rather than a validated photographic improvement.

major comments (4)
  1. [Sec. 4.3, Eqs. (4)-(5); Sec. 7.2, Tables 2-3] The suggestion labels y_s and adjustment labels y_a are computed from CCQA scores, and CCQA is trained on GAICv2 cropping annotations (Sec. 5). Consequently, all objective metrics in Tables 2-3 report how well CPAM reproduces CCQA's argmax and threshold decisions; they do not by themselves establish that the suggested poses improve composition for human viewers. The CCQA generalization test on CPC (Table 4) uses the same crop-ranking paradigm, not full-frame Street View views, so it does not close the transfer gap. Please add an independent human-preference validation of CCQA in the target domain, or a substantially larger user study that directly tests the full SPAS output.
  2. [Sec. 7.3, Table 5] The user study is too small and under-reported to carry the human-validation burden. With 25 students, 100 adjustment pairs and 50 suggestion pairs, no confidence intervals, no inter-rater agreement, and 27% of suggested adjustments judged worse, the evidence is suggestive but not conclusive. Please report per-condition counts, agreement statistics, and statistical tests, and analyze the failure cases to show that the model is not systematically harming composition when it errs.
  3. [Sec. 4.3, Eq. (4)] The adaptive threshold N=25% directly determines the binary suggestion labels for the whole dataset, but the choice is deferred to a Supplementary Material that is not included in the submission. Since the threshold is a free parameter of the labeling pipeline, the paper should report a sensitivity analysis over N and the resulting label distribution, or include the promised supplementary analysis.
  4. [Sec. 4.1, after Eq. (3)] The formula M = (360/Δθ) × (180/Δφ) gives 2,592 candidates at Δθ=Δφ=5°, yet the text reports an average of M=81 candidates after applying Content Preservation and Local Redundancy. Please define the exact candidate-generation algorithm and the resulting M distribution, because the label generation in Eq. (5) depends on the candidate set composition.
minor comments (5)
  1. [Sec. 4.1, Eq. (2) and Fig. 2] The sign conventions for θ and φ should be stated explicitly; the current notation switches between degrees and radians without comment.
  2. [Tables 2 and 3] The two 'IoU↑' columns are ambiguous; please label them clearly as IoU on true-positive cases and IoU on all predicted adjustment cases.
  3. [Figs. 3 and 6] Several symbols in the pipeline figures are garbled (e.g., question marks in place of variables); they should be replaced with proper mathematical notation.
  4. [Sec. 1, contributions] The claims of being 'the first' SPAS system and the first dataset with camera-pose information should be checked against adjacent work on view recommendation and view planning; if prior work exists, the novelty claim should be qualified.
  5. [Sec. 7 and Supplementary references] Several details are deferred to a Supplementary Material that is not present in the reviewed version (e.g., N sensitivity, spherical IoU computation, dataset statistics); these should be included or summarized in the main text.

Circularity Check

1 steps flagged · score 6.0 of 10

CPAM's headline objective results measure agreement with CCQA pseudo-labels, so the composition-improvement claim is circular with respect to its own labeler.

  1. fitted input called prediction [Sec. 4.3, Eqs. (4)-(5); evaluated in Sec. 7.2]
    "Given an initial view I_init ... we use the CCQA model h(·) to assign numerical composition quality ratings to views: s = h(I). ... y_a = (theta_best, phi_best) - (theta_init, phi_init), if y_s = 1 ... where (theta_best, phi_best) represents the camera pose of the candidate view with the highest composition quality score."

    The adjustment and suggestion labels that CPAM is trained and tested on are generated by CCQA: Eq. (5) sets y_a to the displacement toward the CCQA-highest-scoring candidate, and Eq. (4) sets y_s by comparing the CCQA score of the initial view to a CCQA-score threshold. All objective metrics in Sec. 7.2 (AUC, cosine similarity, MAE, IoU) are computed against these same CCQA-derived labels. Hence 'improvement' in those numbers is defined as reproducing CCQA's argmax/ranking; it does not independently establish that the suggested pose produces a better composition. CCQA itself is trained on GAICv2 cropping annotations, and its external CPC test is also a cropping benchmark, so the transfer to full-frame Street View views is not validated by the objective section. The small user study (Sec.

full rationale

The circularity is confined to the objective evaluation chain. CPAM is trained and scored on labels that are themselves produced by CCQA (Eqs. 4-5), so the Sec. 7.2 AUC/CS/MAE/IoU numbers demonstrate that CPAM can imitate CCQA's view ranking, not that following the suggestions improves composition relative to human preference. This is a partial, definitional circularity: the target of prediction is defined by the scoring function used in evaluation. The paper does supply some non-circular evidence: CCQA is trained on human crop scores from GAICv2 and shows competitive top-k accuracy on the external CPC cropping benchmark, and a small human study (Sec. 7.3) directly rates 64% of adjustments as better. These prevent the score from being 8-10, but the central 'improvement' claim in the objective section is still self-referential because neither external benchmark tests full-frame Street View composition. No self-citation chain is involved. Overall score 6.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central derivation rests on a chain of assumptions: geometric projection is standard, but the key load-bearing assumption is that CCQA scores trained on GAICv2 transfer to full-frame Street View views and define what 'good composition' means. The pseudo-labeling step converts that assumption into ground truth for CPAM, so all objective metrics inherit it. Several implementation parameters, including N=25%, step size 5 degrees, alpha=0.1, M=2, and unreported FOV and prompt length, are chosen by hand or by ablation on the test split.

free parameters (6)
  • N (adaptive threshold percentile) = 25%
    Sets tau_i as the Nth highest candidate CCQA score and determines y_s for every scene. Chosen by hand in Sec 4.3 with no sensitivity analysis.
  • alpha (CCQA loss balance) = 0.1
    Balances L3 in Eq. 13; set by hand without reported tuning or ablations.
  • View sampling step sizes = 5 degrees for both delta_theta and delta_phi
    Defines the candidate view grid and the adjustment granularity; adopted from reference [37] and not validated on PCARD.
  • Number of MoE experts M = 2
    Selected by ablation on the PCARD test split (Table 2), so the final test numbers are not fully held-out.
  • Virtual camera field of view = not reported
    The perspective projection in Sec 4.1 requires a field of view to render views, but it is never specified, making exact dataset reproduction impossible.
  • CCQA prompt length L and temperature sigma = not reported
    The learnable prompt context length in Eq. 6 and the softmax temperature in Eq. 7 are left unspecified, affecting pseudo-label quality.
assumptions (6)
  • standard math Perspective projection from a sphere-centered pinhole camera maps yaw and pitch to a planar view.
    Eq. 2 and Sec 4.1 use this mapping to generate candidate views from equirectangular panoramas.
  • domain assumption Composition quality of full-frame views can be scored by a model trained on human cropping annotations (GAICv2).
    CCQA is trained only on GAICv2 and used to label all PCARD views; transfer to Street View scenes is assumed, with CPC generalization as indirect evidence.
  • domain assumption The best composition of a scene is the candidate view with the highest CCQA score.
    Eq. 4-5 define target angles via the argmax CCQA score; the objective evaluation inherits this definition.
  • domain assumption Roll angle can be fixed at zero without harming composition improvement.
    Sec 3 reduces the pose space from three angles to two; no experiment tests whether roll changes would help.
  • domain assumption The overlap constraint (Eq. 3) preserves the photographer's intended subject.
    The threshold lambda in [0.5, 1) is never fixed, and content preservation is asserted rather than measured.
  • domain assumption CCQA pseudo-labels are accurate enough to serve as ground truth for CPAM training.
    The entire learning pipeline depends on this; without human labels on PCARD, label noise is unmeasured.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Smart Point-and-Shoot Photography." pith.science (2026). https://pith.science/paper/PZHAULB7

@misc{pith2026250503638,
  author       = {Pith},
  title        = {Pith review of: Towards Smart Point-and-Shoot Photography},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PZHAULB7}},
  note         = {Machine review of arXiv:2505.03638}
}
read the original abstract

Hundreds of millions of people routinely take photos using their smartphones as point and shoot (PAS) cameras, yet very few would have the photography skills to compose a good shot of a scene. While traditional PAS cameras have built-in functions to ensure a photo is well focused and has the right brightness, they cannot tell the users how to compose the best shot of a scene. In this paper, we present a first of its kind smart point and shoot (SPAS) system to help users to take good photos. Our SPAS proposes to help users to compose a good shot of a scene by automatically guiding the users to adjust the camera pose live on the scene. We first constructed a large dataset containing 320K images with camera pose information from 4000 scenes. We then developed an innovative CLIP-based Composition Quality Assessment (CCQA) model to assign pseudo labels to these images. The CCQA introduces a unique learnable text embedding technique to learn continuous word embeddings capable of discerning subtle visual quality differences in the range covered by five levels of quality description words {bad, poor, fair, good, perfect}. And finally we have developed a camera pose adjustment model (CPAM) which first determines if the current view can be further improved and if so it outputs the adjust suggestion in the form of two camera pose adjustment angles. The two tasks of CPAM make decisions in a sequential manner and each involves different sets of training samples, we have developed a mixture-of-experts model with a gated loss function to train the CPAM in an end-to-end manner. We will present extensive results to demonstrate the performances of our SPAS system using publicly available image composition datasets.

Figures

Figures reproduced from arXiv: 2505.03638 by the authors.

Figure 1
Figure 1. Given a view composed by the user, our Smart Point [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Geometric explanation of the relationship between ERP [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. The overview of our method. Using perspective pro [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: A multi-angle view generation method based on the [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]
Figure 5
Figure 5. Figure 5: CLIP-based Composition Quality Assessment model. [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Camera Pose Adjustment model. The adjustment predictor has a regression head respectively predicting the variables ∆θ and ∆φ. A key challenge is the sequential dependency between these tasks - the adjustment prediction is only meaning￾ful when the suggestion predictor …
Figure 7
Figure 7. Figure 7: Qualitative examples. Each pair shows the original im [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 33 canonical work pages

  1. [1]

    Adaptive fractional dilated convo- lution network for image aesthetics assessment

    Qiuyu Chen, Wei Zhang, Ning Zhou, Peng Lei, Yi Xu, Yu Zheng, and Jianping Fan. Adaptive fractional dilated convo- lution network for image aesthetics assessment. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 14114–14123, 2020. 3

    work page 2020

  2. [2]

    Quantitative analysis of automatic image cropping algorithms: A dataset and comparative study

    Yi-Ling Chen, Tzu-Wei Huang, Kai-Han Chang, Yu-Chen Tsai, Hwann-Tzong Chen, and Bing-Yu Chen. Quantitative analysis of automatic image cropping algorithms: A dataset and comparative study. In 2017 IEEE winter conference on applications of computer vision (WACV), pages 226–234. IEEE, 2017. 2, 3

    work page 2017

  3. [3]

    An experience- based direct generation approach to automatic image crop- ping

    Casper L Christensen and Aneesh Vartakavi. An experience- based direct generation approach to automatic image crop- ping. IEEE Access, 9:107600–107610, 2021. 2

    work page 2021

  4. [4]

    Au- tomatic image cropping using visual composition, boundary simplicity and content preservation models

    Chen Fang, Zhe Lin, Radomir Mech, and Xiaohui Shen. Au- tomatic image cropping using visual composition, boundary simplicity and content preservation models. In Proceedings of the 22nd ACM international conference on Multimedia , pages 1105–1108, 2014. 2, 3

    work page 2014

  5. [5]

    Arucomni: detection of highly reliable fiducial markers in panoramic images

    Jaouad Hajjami, Jordan Caracotte, Guillaume Caron, and Thibault Napoleon. Arucomni: detection of highly reliable fiducial markers in panoramic images. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pages 634–635, 2020. 4

    work page 2020

  6. [6]

    Identity mappings in deep residual networks

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Identity mappings in deep residual networks. In Computer Vision–ECCV 2016: 14th European Conference, Amster- dam, The Netherlands, October 11–14, 2016, Proceedings, Part IV 14, pages 630–645. Springer, 2016. 7

    work page 2016

  7. [7]

    Composing photos like a photographer

    Chaoyi Hong, Shuaiyuan Du, Ke Xian, Hao Lu, Zhiguo Cao, and Weicai Zhong. Composing photos like a photographer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7057–7066, 2021. 3

    work page 2021

  8. [8]

    Learning subject-aware cropping by outpainting professional photos

    James Hong, Lu Yuan, Micha ¨el Gharbi, Matthew Fisher, and Kayvon Fatahalian. Learning subject-aware cropping by outpainting professional photos. In Proceedings of the AAAI Conference on Artificial Intelligence , pages 2175– 2183, 2024

    work page 2024

Show all 37 references
  1. [9]

    Re- thinking image cropping: Exploring diverse compositions from global views

    Gengyun Jia, Huaibo Huang, Chaoyou Fu, and Ran He. Re- thinking image cropping: Exploring diverse compositions from global views. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 2446–2455, 2022. 3

    2022

  2. [10]

    Adam: A method for stochastic opti- mization

    Diederik P Kingma. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980, 2014. 7

    2014 arXiv

  3. [11]

    A2- rl: Aesthetics aware reinforcement learning for image crop- ping

    Debang Li, Huikai Wu, Junge Zhang, and Kaiqi Huang. A2- rl: Aesthetics aware reinforcement learning for image crop- ping. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 8193–8201, 2018. 3

    2018

  4. [12]

    Fast a3rl: Aesthetics-aware adversarial reinforcement learn- ing for image cropping

    Debang Li, Huikai Wu, Junge Zhang, and Kaiqi Huang. Fast a3rl: Aesthetics-aware adversarial reinforcement learn- ing for image cropping. IEEE Transactions on Image Pro- cessing, 28(10):5105–5120, 2019. 3

    2019

  5. [13]

    Composing good shots by exploiting mutual relations

    Debang Li, Junge Zhang, Kaiqi Huang, and Ming-Hsuan Yang. Composing good shots by exploiting mutual relations. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4213–4222, 2020. 3

    2020

  6. [14]

    Composition-aware image aesthetics assess- ment

    Dong Liu, Rohit Puri, Nagendra Kamath, and Subhabrata Bhattacharya. Composition-aware image aesthetics assess- ment. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, pages 3569–3578, 2020. 3

    2020

  7. [15]

    Beyond image borders: Learn- ing feature extrapolation for unbounded image composition

    Xiaoyu Liu, Ming Liu, Junyi Li, Shuai Liu, Xiaotao Wang, Lei Lei, and Wangmeng Zuo. Beyond image borders: Learn- ing feature extrapolation for unbounded image composition. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 13023–13032, 2023. 3

    2023

  8. [16]

    Weakly supervised real-time image cropping based on aesthetic dis- tributions

    Peng Lu, Jiahui Liu, Xujun Peng, and Xiaojie Wang. Weakly supervised real-time image cropping based on aesthetic dis- tributions. In Proceedings of the 28th ACM International Conference on Multimedia, pages 120–128, 2020. 3

    2020

  9. [17]

    A-lamp: Adap- tive layout-aware multi-patch deep convolutional neural net- work for photo aesthetic assessment

    Shuang Ma, Jing Liu, and Chang Wen Chen. A-lamp: Adap- tive layout-aware multi-patch deep convolutional neural net- work for photo aesthetic assessment. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 4535–4544, 2017. 3

    2017

  10. [18]

    Composition-preserving deep photo aesthetics assessment

    Long Mai, Hailin Jin, and Feng Liu. Composition-preserving deep photo aesthetics assessment. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 497–506, 2016. 3

    2016

  11. [19]

    Composition-guided neural network for image cropping aesthetic assessment

    Shijia Ni, Feng Shao, Xiongli Chai, Hangwei Chen, and Yo- Sung Ho. Composition-guided neural network for image cropping aesthetic assessment. IEEE Transactions on Multi- media, 25:6836–6851, 2022. 3

    2022

  12. [20]

    Transview: Inside, outside, and across the cropping view boundaries

    Zhiyu Pan, Zhiguo Cao, Kewei Wang, Hao Lu, and Weicai Zhong. Transview: Inside, outside, and across the cropping view boundaries. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 4218–4227,

  13. [21]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. In International conference on machine learning, ...

    2021

  14. [22]

    Edge detection using moore neighborhood

    Pratibha Sharma, Manoj Diwakar, and Niranjan Lal. Edge detection using moore neighborhood. International Journal of Computer Applications, 61(3), 2013. 4

    2013

  15. [23]

    Spatial-semantic collaborative cropping for user generated content

    Yukun Su, Yiwen Cao, Jingliang Deng, Fengyun Rao, and Qingyao Wu. Spatial-semantic collaborative cropping for user generated content. In Proceedings of the AAAI Con- ference on Artificial Intelligence, pages 4988–4997, 2024. 2, 3

    2024

  16. [24]

    Image cropping with composition and saliency aware aes- thetic score map

    Yi Tu, Li Niu, Weijie Zhao, Dawei Cheng, and Liqing Zhang. Image cropping with composition and saliency aware aes- thetic score map. In Proceedings of the AAAI conference on artificial intelligence, pages 12104–12111, 2020

    2020

  17. [25]

    Image cropping with spatial-aware feature and rank consistency

    Chao Wang, Li Niu, Bo Zhang, and Liqing Zhang. Image cropping with spatial-aware feature and rank consistency. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 10052–10061, 2023. 8

    2023

  18. [26]

    Good view hunting: Learning photo composition from dense view pairs

    Zijun Wei, Jianming Zhang, Xiaohui Shen, Zhe Lin, Radomir Mech, Minh Hoai, and Dimitris Samaras. Good view hunting: Learning photo composition from dense view pairs. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 5437–5446, 2018. 2, 3, 7, 8

    2018

  19. [27]

    Q-align: Teaching lmms for visual scoring via discrete text-defined levels

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Liang Liao, Chunyi Li, Yixuan Gao, Annan Wang, Erli Zhang, Wenxiu Sun, et al. Q-align: Teaching lmms for visual scoring via discrete text-defined levels. arXiv preprint arXiv:2312.17090, 2023. 5

    2023 arXiv

  20. [28]

    Learning the change for automatic image cropping

    Jianzhou Yan, Stephen Lin, Sing Bing Kang, and Xiaoou Tang. Learning the change for automatic image cropping. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 971–978, 2013. 2, 3

    2013

  21. [29]

    Focusing on your subject: Deep subject-aware image composition recommendation net- works

    Guo-Ye Yang, Wen-Yang Zhou, Yun Cai, Song-Hai Zhang, and Fang-Lue Zhang. Focusing on your subject: Deep subject-aware image composition recommendation net- works. Computational Visual Media, 9(1):87–107, 2023. 2, 3

    2023

  22. [30]

    Aesthetic image cropping meets vlp: Enhancing good while reducing bad

    Quan Yuan, Leida Li, and Pengfei Chen. Aesthetic image cropping meets vlp: Enhancing good while reducing bad. Journal of Visual Communication and Image Representation, page 104316, 2024. 3

    2024

  23. [31]

    Reliable and efficient image cropping: A grid anchor based approach

    Hui Zeng, Lida Li, Zisheng Cao, and Lei Zhang. Reliable and efficient image cropping: A grid anchor based approach. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 5949–5957, 2019. 2, 3

    2019

  24. [32]

    Grid anchor based image cropping: A new benchmark and an efficient model

    Hui Zeng, Lida Li, Zisheng Cao, and Lei Zhang. Grid anchor based image cropping: A new benchmark and an efficient model. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(3):1304–1319, 2020. 2, 3, 5, 7, 8

    2020

  25. [33]

    Image composition assessment with saliency-augmented multi-pattern pooling

    Bo Zhang, Li Niu, and Liqing Zhang. Image composition assessment with saliency-augmented multi-pattern pooling. arXiv preprint arXiv:2104.03133, 2021. 3

    2021 arXiv

  26. [34]

    Blind image quality assessment via vision- language correspondence: A multitask learning perspective

    Weixia Zhang, Guangtao Zhai, Ying Wei, Xiaokang Yang, and Kede Ma. Blind image quality assessment via vision- language correspondence: A multitask learning perspective. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 14071–14081, 2023. 5

    2023

  27. [35]

    Spherical criteria for fast and accurate 360 object detection

    Pengyu Zhao, Ansheng You, Yuanxing Zhang, Jiaying Liu, Kaigui Bian, and Yunhai Tong. Spherical criteria for fast and accurate 360 object detection. In Proceedings of the AAAI Conference on Artificial Intelligence , pages 12959–12966,

  28. [36]

    Aesthetic-guided outward im- age cropping

    Lei Zhong, Feng-Heng Li, Hao-Zhi Huang, Yong Zhang, Shao-Ping Lu, and Jue Wang. Aesthetic-guided outward im- age cropping. ACM Transactions on Graphics (TOG), 40(6): 1–13, 2021. 3

    2021

  29. [37]

    Stable viewport-based unsupervised compressed 360◦ video quality enhancement

    Zizhuang Zou, Mao Ye, Xue Li, Luping Ji, and Ce Zhu. Stable viewport-based unsupervised compressed 360◦ video quality enhancement. IEEE Transactions on Broadcasting,

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.