A bi-directional evaluation-based approach for image retargeting quality assessment

Oliveira, Saulo A. F.; Alves, Shara S. A.; Gomes, João P. P.; Neto, Ajalmar R. da Rocha

doi:10.1016/j.cviu.2017.11.011

Cited by 12 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Table II, we present a comparison between our best result and other IRQAs in the literature. The other IRQAs scores were taken directly from [15]. Concerning LLC, SRCC, and RMSE, AEIMPS achieved an intermediate rank compared to the others.…”

Section: B Results and Discussionmentioning

confidence: 99%

“…Many authors have proposed Image Retargeting Quality Algorithms (IRQAs) to overcome such a drawback [11]- [14]. Usually, such algorithms create a pixel matching mapping between the original image to the retargeting one, indicating a degree of content preservation [15]. Then, after applying some similarity criterion or measure of distance [16], one can recall the retargeting quality based on content matching and, perhaps, content relevance.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

AEIMPS: Deep Autoencoder for Image Retargeting Quality Assessment

Carvalho¹,

Oliveira²

2022

Anais Estendidos Do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)

View full text Add to dashboard Cite

Evaluating retargeting image operators is a subjective task and, therefore, challenging to execute without human interference. Image Retargeting Quality Algorithms execute this task, giving some score to the retargeted image and, usually, trying to get a result similar to a human opinion since humans generally agree with each other on the quality of a resized image. Therefore, we propose an Autoencoder-based IRQA named AutoEncoder Information MaP Similarity (AEIMPS) to address this task using the NVAE architecture. In our experiments, besides the retargeting ratio, we use the latent space and the reconstructed image in the IRQA. AIEMPS achieved an average performance compared to other IRQAs in the literature.

show abstract

Section: B Results and Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

AEIMPS: Deep Autoencoder for Image Retargeting Quality Assessment

Carvalho¹,

Oliveira²

2022

Anais Estendidos Do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)

View full text Add to dashboard Cite

show abstract

“…We published the results of this thesis in Neurocomputing [4], [5]. We also would like to highlight that Saulo A. F. Oliveira also contributed as the main author in Computer Vision and Image Understanding [11] and co-authored two other works in Applied Soft Computing [12] and Soft Computing [13]. All above mentioned journals have Qualis A1.…”

Section: Publicationsmentioning

confidence: 88%

On Model Complexity Reduction in Instance-Based Learners

Oliveira

Neto²,

Gomes

2022

Anais Estendidos Do XXXV Conference on Graphics, Patterns and Images (SIBGRAPI Estendido 2022)

View full text Add to dashboard Cite

Instance-based learners habitually adopt instance selection techniques to reduce complexity and avoid overfitting. Such learners’ most recent and well-known formulations seek to impose some sparsity in their training and prediction structure alongside regularization to meet such a result. Due to the variety of such instance-based learners, we will draw attention to the Least-Squares Support Vector Machines and Minimal Learning Machines because they embody additional information beyond the stored instances to perform predictions. Later, this thesis proposes variants constraining candidate solutions within a specific functional space where we avoid overfitting and reduce model complexity. The central core of such variants is related to penalizing samples with a specific condition during learning. For regressors, we adopted strategies based on random and observed linearity conditions related to the data. At the same time, we borrowed definitions from the computer vision field for classification tasks to derive a concept we call the classcorner relationship (in which we designed an instance selection algorithm). In the Least-Squares Support Vector Machines context, this thesis follows the pruning fashion by adopting the samples that share such a class-corner relationship. As for the Minimal Learning Machine model, this thesis introduces a new proposal called the Lightweight Minimal Learning Machine, a faster model for out-of-sample prediction due to the reduced number of computations inherent in the original proposal’s multilateration process. Another remarkable feature is that it derives a unique solution when other formulations rely on overdetermined systems.

show abstract

“…Jiang et al [30] focused on learning a sparse representation of an image that contains distortion sensitive features. Oliveira et al [31] measured the loss of relevant content and visual artifacts created in retargeted images in a bi-directional approach. Zhang et al [32] analyzed in three levels including region-level, patch-level and pixel-level and made effort in detecting deformation inconsistency.…”

Section: Related Workmentioning

confidence: 99%

A Content-Aware Image Retargeting Quality Assessment Method Using Foreground and Global Measurement

Guo

Jin

2019

IEEE Access

View full text Add to dashboard Cite

Image retargeting methods aim to minimize the perceptual loss while changing sizes and aspect ratios of images. Since optimal retargeting methods for different images are generally not the same, the image retargeting quality assessment (IRQA) becomes a meaningful task. This paper proposes a content-aware image retargeting quality assessment method using foreground and global measurement to achieve better performance. In our proposed method, images are first divided into two categories according to the foreground object detection result, and then different corresponding measurements are designed for them. For those with obvious foreground object, both foreground and global measurement are applied. For others, only global measurement is conducted. Foreground measurement includes two complementary features: the high-level semantic similarity feature and the low-level size ratio feature. Global measurement includes another two features: an improved aspect ratio similarity (ARS) feature and edge group similarity (EGS) feature. Two public databases, i.e., the RetargetMe and CUHK, have been evaluated, and experimental results demonstrate that our method is quite effective, and it also provides state-of-the-art performance in the IRQA. a INDEX TERMS Image retargeting quality assessment, image quality assessment, foreground object quality measurement, semantic similarity feature.

show abstract

A bi-directional evaluation-based approach for image retargeting quality assessment

Cited by 12 publications

References 16 publications

AEIMPS: Deep Autoencoder for Image Retargeting Quality Assessment

AEIMPS: Deep Autoencoder for Image Retargeting Quality Assessment

On Model Complexity Reduction in Instance-Based Learners

A Content-Aware Image Retargeting Quality Assessment Method Using Foreground and Global Measurement

Contact Info

Product

Resources

About