Jessica de Souza
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Jessica de Souza has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Edward Jay Wang
University of California San Diego
5 shared publications - Kristina Chamberlain
Univ Calif San Diego, Div Extended Studies, La Jolla, CA USA
3 shared publications - Ariana Talai
The Design Lab, University of California San Diego, La Jolla, California, USA
2 shared publications - Jessica Maria Echterhoff
Univ Calif San Diego, Dept Comp Sci & Engn, La Jolla, CA USA
2 shared publications - Tanya Punater
Herbert Wertheim School of Public Health and Human Longevity Science, University of California San Diego, La Jolla, California, USA
2 shared publications - Varun Kumar Viswanath
Department of Electrical and Computer Engineering, Jacobs School of Engineering, University of California, San Diego, USA
2 shared publications
Latest Publications
- Open Access
- 7 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Understanding the Challenges and Design Opportunities of Using Voice Assistants to Support Postpartum Mothers in Brazil
- 7 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Open AccessUnderstanding the Challenges and Design Opportunities of Using Voice Assistants to Support Postpartum Mothers in Brazil
- Open Access
- 24 June 2024
- Published by JMIR Publications Inc. in JMIR AI
- Vol. 3, e54798
Augmenting Telepostpartum Care With Vision-Based Detection of Breastfeeding-Related Conditions: Algorithm Development and Validation
Background: Breastfeeding benefits both the mother and infant and is a topic of attention in public health. After childbirth, untreated medical conditions or lack of support lead many mothers to discontinue breastfeeding. For instance, nipple damage and mastitis affect 80% and 20% of US mothers, respectively. Lactation consultants (LCs) help mothers with breastfeeding, providing in-person, remote, and hybrid lactation support. LCs guide, encourage, and find ways for mothers to have a better experience breastfeeding. Current telehealth services help mothers seek LCs for breastfeeding support, where images help them identify and address many issues. Due to the disproportional ratio of LCs and mothers in need, these professionals are often overloaded and burned out. Objective: This study aims to investigate the effectiveness of 5 distinct convolutional neural networks in detecting healthy lactating breasts and 6 breastfeeding-related issues by only using red, green, and blue images. Our goal was to assess the applicability of this algorithm as an auxiliary resource for LCs to identify painful breast conditions quickly, better manage their patients through triage, respond promptly to patient needs, and enhance the overall experience and care for breastfeeding mothers. Methods: We evaluated the potential for 5 classification models to detect breastfeeding-related conditions using 1078 breast and nipple images gathered from web-based and physical educational resources. We used the convolutional neural networks Resnet50, Visual Geometry Group model with 16 layers (VGG16), InceptionV3, EfficientNetV2, and DenseNet169 to classify the images across 7 classes: healthy, abscess, mastitis, nipple blebs, dermatosis, engorgement, and nipple damage by improper feeding or misuse of breast pumps. We also evaluated the models’ ability to distinguish between healthy and unhealthy images. We present an analysis of the classification challenges, identifying image traits that may confound the detection model. Results: The best model achieves an average area under the receiver operating characteristic curve of 0.93 for all conditions after data augmentation for multiclass classification. For binary classification, we achieved, with the best model, an average area under the curve of 0.96 for all conditions after data augmentation. Several factors contributed to the misclassification of images, including similar visual features in the conditions that precede other conditions (such as the mastitis spectrum disorder), partially covered breasts or nipples, and images depicting multiple conditions in the same breast. Conclusions: This vision-based automated detection technique offers an opportunity to enhance postpartum care for mothers and can potentially help alleviate the workload of LCs by expediting decision-making processes.
- 24 June 2024
- Published by JMIR Publications Inc. in JMIR AI
- Vol. 3, e54798
Open AccessAugmenting Telepostpartum Care With Vision-Based Detection of Breastfeeding-Related Conditions: Algorithm Development and Validation
Background: Breastfeeding benefits both the mother and infant and is a topic of attention in public health. After childbirth, untreated medical conditions or lack of support lead many mothers to discontinue breastfeeding. For instance, nipple damage and mastitis affect 80% and 20% of US mothers, respectively. Lactation consultants (LCs) help mothers with breastfeeding, providing in-person, remote, and hybrid lactation support. LCs guide, encourage, and find ways for mothers to have a better experience breastfeeding. Current telehealth services help mothers seek LCs for breastfeeding support, where images help them identify and address many issues. Due to the disproportional ratio of LCs and mothers in need, these professionals are often overloaded and burned out. Objective: This study aims to investigate the effectiveness of 5 distinct convolutional neural networks in detecting healthy lactating breasts and 6 breastfeeding-related issues by only using red, green, and blue images. Our goal was to assess the applicability of this algorithm as an auxiliary resource for LCs to identify painful breast conditions quickly, better manage their patients through triage, respond promptly to patient needs, and enhance the overall experience and care for breastfeeding mothers. Methods: We evaluated the potential for 5 classification models to detect breastfeeding-related conditions using 1078 breast and nipple images gathered from web-based and physical educational resources. We used the convolutional neural networks Resnet50, Visual Geometry Group model with 16 layers (VGG16), InceptionV3, EfficientNetV2, and DenseNet169 to classify the images across 7 classes: healthy, abscess, mastitis, nipple blebs, dermatosis, engorgement, and nipple damage by improper feeding or misuse of breast pumps. We also evaluated the models’ ability to distinguish between healthy and unhealthy images. We present an analysis of the classification challenges, identifying image traits that may confound the detection model. Results: The best model achieves an average area under the receiver operating characteristic curve of 0.93 for all conditions after data augmentation for multiclass classification. For binary classification, we achieved, with the best model, an average area under the curve of 0.96 for all conditions after data augmentation. Several factors contributed to the misclassification of images, including similar visual features in the conditions that precede other conditions (such as the mastitis spectrum disorder), partially covered breasts or nipples, and images depicting multiple conditions in the same breast. Conclusions: This vision-based automated detection technique offers an opportunity to enhance postpartum care for mothers and can potentially help alleviate the workload of LCs by expediting decision-making processes.
- Open Access
- 10 June 2024
- Published by Frontiers Media SA in Frontiers in Virtual Reality
- Vol. 5, 1368721
Visual perception and user satisfaction in video see-through head-mounted displays: a mixed-methods evaluation
Objective: Our study addresses the challenges limiting the adoption of Extended Reality (XR) Head-Mounted Displays (HMDs), mainly focusing on device quality and cybersickness. We aim to investigate the impact of hardware and software on user experience and task performance while wearing Video See-Through (VST) HMDs. We employ a novel methodology designed to bridge the gaps identified in previous research.Methods: This study uses a convergent mixed-methods approach, combining qualitative and quantitative data in a within-subjects evaluation involving 20 participants. This comprehensive evaluation examines visual perception, visual quality, and user experience through a range of tasks. Usability, comfort, and cybersickness are assessed, with insights derived from both user performance metrics and subjective measures collected through in-depth interviews and comments. The study includes three distinct HMDs—two prototypes (PD1 and PD2) and one commercial device (CD1)—to provide a broad analysis of the technology.Findings: Our findings reveal that while participants were generally satisfied with VST mode, their preferences varied across devices. CD1 was preferred for its realistic color representation and superior reading task performance due to its high-resolution display and camera. However, visual disturbances and temporal issues differed across devices, with CD1 exhibiting fewer artifacts when stationary but showing more disturbances when participants were moving. Participants found PD1 and PD2 more comfortable for extended use and fewer cybersickness symptoms, but they highlighted color and display resolution issues. These variations underscore the importance of considering both qualitative and quantitative measures in HMD evaluations.Conclusion: This mixed-methods evaluation emphasizes the limitations of relying solely on visual perception performance measures for VST HMDs. By integrating both quantitative and qualitative insights, we offer a more detailed evaluation framework to identify design flaws and user experience issues that quantitative metrics alone might miss. This methodology contributes to the field by illustrating how a mixed-methods approach provides a broader perspective on XR technology, guiding future improvements and enhancing VST adoption in future applications.
- 10 June 2024
- Published by Frontiers Media SA in Frontiers in Virtual Reality
- Vol. 5, 1368721
Open AccessVisual perception and user satisfaction in video see-through head-mounted displays: a mixed-methods evaluation
Objective: Our study addresses the challenges limiting the adoption of Extended Reality (XR) Head-Mounted Displays (HMDs), mainly focusing on device quality and cybersickness. We aim to investigate the impact of hardware and software on user experience and task performance while wearing Video See-Through (VST) HMDs. We employ a novel methodology designed to bridge the gaps identified in previous research.Methods: This study uses a convergent mixed-methods approach, combining qualitative and quantitative data in a within-subjects evaluation involving 20 participants. This comprehensive evaluation examines visual perception, visual quality, and user experience through a range of tasks. Usability, comfort, and cybersickness are assessed, with insights derived from both user performance metrics and subjective measures collected through in-depth interviews and comments. The study includes three distinct HMDs—two prototypes (PD1 and PD2) and one commercial device (CD1)—to provide a broad analysis of the technology.Findings: Our findings reveal that while participants were generally satisfied with VST mode, their preferences varied across devices. CD1 was preferred for its realistic color representation and superior reading task performance due to its high-resolution display and camera. However, visual disturbances and temporal issues differed across devices, with CD1 exhibiting fewer artifacts when stationary but showing more disturbances when participants were moving. Participants found PD1 and PD2 more comfortable for extended use and fewer cybersickness symptoms, but they highlighted color and display resolution issues. These variations underscore the importance of considering both qualitative and quantitative measures in HMD evaluations.Conclusion: This mixed-methods evaluation emphasizes the limitations of relying solely on visual perception performance measures for VST HMDs. By integrating both quantitative and qualitative insights, we offer a more detailed evaluation framework to identify design flaws and user experience issues that quantitative metrics alone might miss. This methodology contributes to the field by illustrating how a mixed-methods approach provides a broader perspective on XR technology, guiding future improvements and enhancing VST adoption in future applications.
- Open Access
- 2 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-7
LCBuddy: Towards a Smartphone-based Self-Assessment Tool for Postpartum Patients With Breast Pain
Telehealth has significantly enhanced various medical sectors, notably in postpartum care, by increasing breastfeeding rates and reducing postpartum depression. However, issues in patient management and triage, particularly in high-demand telehealth lactation consultations, lead to LC burnout, unnecessary costs and diagnostic delays. To address these challenges, this study introduces a pain assessment tool for telehealth lactation consultations, enhancing patient-provider communication through momentary assessment and AI mediation. It features a user interface with image quality assessment and condition classification algorithms, aiding remote diagnosis and triage. The tool guides patients and uses a severity-based feedback system, helping LCs prioritize cases. This method boosts tele-lactation service efficiency, streamlines LC workload, and serves as a model for specialized healthcare triaging. Future work includes refining the feedback system and evaluating the tool in clinical settings to standardize symptom severity grading and estimate its impact, potentially improving expert patient triaging in remote healthcare settings.
- 2 May 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-7
Open AccessLCBuddy: Towards a Smartphone-based Self-Assessment Tool for Postpartum Patients With Breast Pain
Telehealth has significantly enhanced various medical sectors, notably in postpartum care, by increasing breastfeeding rates and reducing postpartum depression. However, issues in patient management and triage, particularly in high-demand telehealth lactation consultations, lead to LC burnout, unnecessary costs and diagnostic delays. To address these challenges, this study introduces a pain assessment tool for telehealth lactation consultations, enhancing patient-provider communication through momentary assessment and AI mediation. It features a user interface with image quality assessment and condition classification algorithms, aiding remote diagnosis and triage. The tool guides patients and uses a severity-based feedback system, helping LCs prioritize cases. This method boosts tele-lactation service efficiency, streamlines LC workload, and serves as a model for specialized healthcare triaging. Future work includes refining the feedback system and evaluating the tool in clinical settings to standardize symptom severity grading and estimate its impact, potentially improving expert patient triaging in remote healthcare settings.
- Open Access
- 22 November 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Augmenting Telepostpartum Care With Vision-Based Detection of Breastfeeding-Related Conditions: Algorithm Development and Validation (Preprint)
BACKGROUND Breastfeeding benefits both the mother and infant and is a topic of attention in public health. After childbirth, untreated medical conditions or lack of support lead many mothers to discontinue breastfeeding. For instance, nipple damage and mastitis affect 80% and 20% of US mothers, respectively. Lactation consultants (LCs) help mothers with breastfeeding, providing in-person, remote, and hybrid lactation support. LCs guide, encourage, and find ways for mothers to have a better experience breastfeeding. Current telehealth services help mothers seek LCs for breastfeeding support, where images help them identify and address many issues. Due to the disproportional ratio of LCs and mothers in need, these professionals are often overloaded and burned out. OBJECTIVE This study aims to investigate the effectiveness of 5 distinct convolutional neural networks in detecting healthy lactating breasts and 6 breastfeeding-related issues by only using red, green, and blue images. Our goal was to assess the applicability of this algorithm as an auxiliary resource for LCs to identify painful breast conditions quickly, better manage their patients through triage, respond promptly to patient needs, and enhance the overall experience and care for breastfeeding mothers. METHODS We evaluated the potential for 5 classification models to detect breastfeeding-related conditions using 1078 breast and nipple images gathered from web-based and physical educational resources. We used the convolutional neural networks Resnet50, Visual Geometry Group model with 16 layers (VGG16), InceptionV3, EfficientNetV2, and DenseNet169 to classify the images across 7 classes: healthy, abscess, mastitis, nipple blebs, dermatosis, engorgement, and nipple damage by improper feeding or misuse of breast pumps. We also evaluated the models’ ability to distinguish between healthy and unhealthy images. We present an analysis of the classification challenges, identifying image traits that may confound the detection model. RESULTS The best model achieves an average area under the receiver operating characteristic curve of 0.93 for all conditions after data augmentation for multiclass classification. For binary classification, we achieved, with the best model, an average area under the curve of 0.96 for all conditions after data augmentation. Several factors contributed to the misclassification of images, including similar visual features in the conditions that precede other conditions (such as the mastitis spectrum disorder), partially covered breasts or nipples, and images depicting multiple conditions in the same breast. CONCLUSIONS This vision-based automated detection technique offers an opportunity to enhance postpartum care for mothers and can potentially help alleviate the workload of LCs by expediting decision-making processes.
- 22 November 2023
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessAugmenting Telepostpartum Care With Vision-Based Detection of Breastfeeding-Related Conditions: Algorithm Development and Validation (Preprint)
BACKGROUND Breastfeeding benefits both the mother and infant and is a topic of attention in public health. After childbirth, untreated medical conditions or lack of support lead many mothers to discontinue breastfeeding. For instance, nipple damage and mastitis affect 80% and 20% of US mothers, respectively. Lactation consultants (LCs) help mothers with breastfeeding, providing in-person, remote, and hybrid lactation support. LCs guide, encourage, and find ways for mothers to have a better experience breastfeeding. Current telehealth services help mothers seek LCs for breastfeeding support, where images help them identify and address many issues. Due to the disproportional ratio of LCs and mothers in need, these professionals are often overloaded and burned out. OBJECTIVE This study aims to investigate the effectiveness of 5 distinct convolutional neural networks in detecting healthy lactating breasts and 6 breastfeeding-related issues by only using red, green, and blue images. Our goal was to assess the applicability of this algorithm as an auxiliary resource for LCs to identify painful breast conditions quickly, better manage their patients through triage, respond promptly to patient needs, and enhance the overall experience and care for breastfeeding mothers. METHODS We evaluated the potential for 5 classification models to detect breastfeeding-related conditions using 1078 breast and nipple images gathered from web-based and physical educational resources. We used the convolutional neural networks Resnet50, Visual Geometry Group model with 16 layers (VGG16), InceptionV3, EfficientNetV2, and DenseNet169 to classify the images across 7 classes: healthy, abscess, mastitis, nipple blebs, dermatosis, engorgement, and nipple damage by improper feeding or misuse of breast pumps. We also evaluated the models’ ability to distinguish between healthy and unhealthy images. We present an analysis of the classification challenges, identifying image traits that may confound the detection model. RESULTS The best model achieves an average area under the receiver operating characteristic curve of 0.93 for all conditions after data augmentation for multiclass classification. For binary classification, we achieved, with the best model, an average area under the curve of 0.96 for all conditions after data augmentation. Several factors contributed to the misclassification of images, including similar visual features in the conditions that precede other conditions (such as the mastitis spectrum disorder), partially covered breasts or nipples, and images depicting multiple conditions in the same breast. CONCLUSIONS This vision-based automated detection technique offers an opportunity to enhance postpartum care for mothers and can potentially help alleviate the workload of LCs by expediting decision-making processes.