Predicting body measures from 2D images using Convolutional Neural Networks

Souza, João Wellington M.; Holanda, Gabriel B.; Ivo, Roberto F.; Alves, Shara S. A.; Silva, Suane P. P. da; Nunes, Virgínia Xavier; Loureiro, Luiz Lannes; Dias-Silva, C. H.; Filho, Pedro Pedrosa Rebouças

doi:10.1109/ijcnn48605.2020.9207330

Cited by 9 publications

(9 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the images had to be taken in front of clean black backgrounds, and participants needed to wear measurement attire and caps (Lin and Wang, 2011). Other research used 2D images to run statistical predictions on measurements without forming 3D models (de Souza et al, 2020;Moses et al, 2013;Xia et al, 2018). To help determine the reference scale in the images, Moses et al (2013) Another way to use the 2D images is to virtually stitch 2D images into 3D models based on texture and silhouette information in the images (Lensch et al, 2001).…”

Section: Literature Reviewmentioning

confidence: 99%

See 1 more Smart Citation

Fashionable sportswear working as a body measurement collecting tool

Xia

West

et al. 2022

IJCST

View full text Add to dashboard Cite

PurposeA two-dimensional (2D) body measurement system was implemented to study the application of sportswear design in measurement garment development. A total of 50 participants were recruited. The basic demographic information and sportswear preference data were collected through a survey to understand consumer preferences and acceptance of the new designs. The body measurements were collected through both the selected 2D measurement system and a commercial three-dimensional scanning booth to evaluate measurement accuracy.Design/methodology/approachFinding the right size has been a long-existing problem for clothing consumers. Size problem is the most common reason for e-commerce returns and adds a high cost for retailers. One possible solution is to offer consumers an easy-to-use method that extracts accurate body measurements to be used for clothing size selection. The purpose of this research is to apply sportswear design elements on measurement garments to see if consumers’ interest in using the 2D measurement system can be increased without influencing the measurement accuracy.FindingsThe results showed that the added design features increased consumers’ interest in using 2D body measuring technology without significantly influencing measurement accuracy.Originality/valueThis research applied sportswear elements to convert a 2D measurement bodysuit to a fashionable clothing product. The solution resolved users’ privacy concerns and increased their acceptance and use of the technology. Other studies have not focused on using aesthetic features to improve the 2D measurement technology.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

Fashionable sportswear working as a body measurement collecting tool

Xia

West

et al. 2022

IJCST

View full text Add to dashboard Cite

show abstract

“…, 2018). de Souza et al. (2020) used Dense Human Pose Estimation (DensePose), a variation of RCNN, to separate body shapes from backgrounds.…”

Section: Related Workmentioning

confidence: 99%

“…Researchers have explored using neural network models to solve fashion-related problems such as fashion parsing, style learning, pose transformation, and fashion recommendation (Cheng et al, 2021;Luce, 2018) (Ayyadevara and Reddy, 2020;Khan et al, 2018). de Souza et al (2020) used Dense Human Pose Estimation (DensePose), a variation of RCNN, to separate body shapes from backgrounds. However, this research still required subjects to wear a minimum of clothing and stand in front of a clean background.…”

Section: Related Workmentioning

confidence: 99%

A 2D image body measurement system developed with neural networks and a color-coded measurement garment

Xia

Istook

et al. 2022

IJCST

View full text Add to dashboard Cite

PurposeTwo-dimensional (2D) measurement technology has become more popular than before, thanks to the widespread availability of smartphones and smart devices. However, most existing 2D body measurement systems have background constraints and may raise privacy concerns. The purpose of this research was to test the idea of designing a 2D measurement system that works with a color-coded measurement garment for background removal and privacy protection. Clothing consumers can use the proposed system for daily apparel shopping purposes.Design/methodology/approachA 2D body measurement system was designed and tested. The system adopted a close-fitted color-coded measurement garment and used neural network models to detect the color-code in the garment area and remove backgrounds. In total, 78 participants were recruited, and the collected data were split into training and testing sets. The training dataset was used to train the neural network and statistical prediction models for the 2D system. The testing dataset was used to compare the performance of the 2D system with a commercial three-dimensional (3D) body scanner.FindingsThe results showed that the color-coded measurement garment worked well with the neural network models to process the images for measurement extraction. The 2D measurement system worked better at close-fitted areas than loose-fitted areas.Originality/valueThis research combined a color-coded measurement garment with neural network models to solve the privacy and background challenges of the 2D body measurement system. Other researchers have never studied this approach.

show abstract

“…Hu et al 19 used adaptive body structure segmentation (ABSS) adaptively to segment the key regions of human body structure, and specific algorithms for different parts of the region to extract feature points. Based on the front and back photos of the subjects, De Souza et al 20 and others used convolutional neural networks and deep learning technology to segment the human body photo semantically and identify feature points, and the actual sizes of the human body were calculated based on the ratio of the actual height to the photo height. In addition, some scholars extracted the sizes by locating the feature points of the human body after extracting the outline of the human body.…”

mentioning

confidence: 99%

Automatic recognition of the neck–shoulder shape based on 2D photos

Zhang

ZHAO

et al. 2022

Textile Research Journal

View full text Add to dashboard Cite

To improve body type classification and automatic identification, this study adopted a method by combining three-dimensional (3D) point-cloud data with two-dimensional (2D) photos for the classification of young mens’ neck–shoulder shape. A total of 180 male college students were measured by photo and 3D body measuring methods to get the main body information. As for the 3D body cloud data, reverse engineering software was used to measure the parameters closely related to the neck–shoulder shape. Five important parameters, including the forward angle, back angle, shoulder angle, the ratio between neck and shoulder width, and the ratio between neck width and thickness were analyzed to divide young men’s neck–shoulder shape into three categories, namely round drop shoulder, forward round neck and wide straight neck, and the classification rules were established. As for the 2D photos (such as front and side photos), the human body contour was extracted to obtain the parameters required for neck–shoulder shape classification by identifying the feature points. The neck–shoulder shape could be automatically recognized, and the accuracy rate is 97.7% to verify that the neck–shoulder shape recognition system is effective. This study proposed a simple neck–shoulder shape automatic recognition method based on the 2D body photos, and provide reference for automatic neck–shoulder shape recognition, 3D modeling and virtual fitting of local features.

show abstract

Predicting body measures from 2D images using Convolutional Neural Networks

Cited by 9 publications

References 17 publications

Fashionable sportswear working as a body measurement collecting tool

Fashionable sportswear working as a body measurement collecting tool

A 2D image body measurement system developed with neural networks and a color-coded measurement garment

Automatic recognition of the neck–shoulder shape based on 2D photos

Contact Info

Product

Resources

About