Image Turing test and its applications on synthetic chest radiographs by using the progressive growing generative adversarial network

Jang, Miso; Bae, Hyun‐Jin; Kim, Minjee; Park, Seo Young; Son, A-Yeon; Choi, Se Jin; Choe, Jooae; Choi, Hye Young; Hwang, Hye Jeon; Noh, Han Na; Seo, Joon Beom; Lee, Sang Min; Kim, Namkug

doi:10.1038/s41598-023-28175-1

Cited by 5 publications

(4 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Embryologists (Group I) and IVF lab technicians (Group II) achieved a higher accuracy than non-experts in distinguishing real from generated embryo images. However, the accuracy rate for embryologists (Group I) is only 55.7%, compared to 61.3% from assessing synthetic gastroscopy images ( Shin et al , 2023 ), and 67.4%, 69.9% from evaluating two sets of generated chest radiographs ( Jang et al , 2023 ). Additionally, in this study, individuals without specialized expertise in embryo imaging performed at a level approximately equivalent to random guessing (50%).…”

Section: Discussionmentioning

confidence: 99%

Generative artificial intelligence to produce high-fidelity blastocyst-stage embryo images

Cao,

Derhaag,

Coonen

et al. 2024

Human Reproduction

View full text Add to dashboard Cite

STUDY QUESTION Can generative artificial intelligence (AI) models produce high-fidelity images of human blastocysts? SUMMARY ANSWER Generative AI models exhibit the capability to generate high-fidelity human blastocyst images, thereby providing substantial training datasets crucial for the development of robust AI models. WHAT IS KNOWN ALREADY The integration of AI into IVF procedures holds the potential to enhance objectivity and automate embryo selection for transfer. However, the effectiveness of AI is limited by data scarcity and ethical concerns related to patient data privacy. Generative adversarial networks (GAN) have emerged as a promising approach to alleviate data limitations by generating synthetic data that closely approximate real images. STUDY DESIGN, SIZE, DURATION Blastocyst images were included as training data from a public dataset of time-lapse microscopy (TLM) videos (n = 136). A style-based GAN was fine-tuned as the generative model. PARTICIPANTS/MATERIALS, SETTING, METHODS We curated a total of 972 blastocyst images as training data, where frames were captured within the time window of 110–120 h post-insemination at 1-h intervals from TLM videos. We configured the style-based GAN model with data augmentation (AUG) and pretrained weights (Pretrained-T: with translation equivariance; Pretrained-R: with translation and rotation equivariance) to compare their optimization on image synthesis. We then applied quantitative metrics including Fréchet Inception Distance (FID) and Kernel Inception Distance (KID) to assess the quality and fidelity of the generated images. Subsequently, we evaluated qualitative performance by measuring the intelligence behavior of the model through the visual Turing test. To this end, 60 individuals with diverse backgrounds and expertise in clinical embryology and IVF evaluated the quality of synthetic embryo images. MAIN RESULTS AND THE ROLE OF CHANCE During the training process, we observed consistent improvement of image quality that was measured by FID and KID scores. Pretrained and AUG + Pretrained initiated with remarkably lower FID and KID values compared to both Baseline and AUG + Baseline models. Following 5000 training iterations, the AUG + Pretrained-R model showed the highest performance of the evaluated five configurations with FID and KID scores of 15.2 and 0.004, respectively. Subsequently, we carried out the visual Turing test, such that IVF embryologists, IVF laboratory technicians, and non-experts evaluated the synthetic blastocyst-stage embryo images and obtained similar performance in specificity with marginal differences in accuracy and sensitivity. LIMITATIONS, REASONS FOR CAUTION In this study, we primarily focused the training data on blastocyst images as IVF embryos are primarily assessed in blastocyst stage. However, generation of an array of images in different preimplantation stages offers further insights into the development of preimplantation embryos and IVF success. In addition, we resized training images to a resolution of 256 × 256 pixels to moderate the computational costs of training the style-based GAN models. Further research is needed to involve a more extensive and diverse dataset from the formation of the zygote to the blastocyst stage, e.g. video generation, and the use of improved image resolution to facilitate the development of comprehensive AI algorithms and to produce higher-quality images. WIDER IMPLICATIONS OF THE FINDINGS Generative AI models hold promising potential in generating high-fidelity human blastocyst images, which allows the development of robust AI models as it can provide sufficient training datasets while safeguarding patient data privacy. Additionally, this may help to produce sufficient embryo imaging training data with different (rare) abnormal features, such as embryonic arrest, tripolar cell division to avoid class imbalances and reach to even datasets. Thus, generative models may offer a compelling opportunity to transform embryo selection procedures and substantially enhance IVF outcomes. STUDY FUNDING/COMPETING INTEREST(S) This study was supported by a Horizon 2020 innovation grant (ERIN, grant no. EU952516) and a Horizon Europe grant (NESTOR, grant no. 101120075) of the European Commission to A.S. and M.Z.E., the Estonian Research Council (grant no. PRG1076) to A.S., and the EVA (Erfelijkheid Voortplanting & Aanleg) specialty program (grant no. KP111513) of Maastricht University Medical Centre (MUMC+) to M.Z.E. TRIAL REGISTRATION NUMBER Not applicable.

show abstract

Section: Discussionmentioning

confidence: 99%

Generative artificial intelligence to produce high-fidelity blastocyst-stage embryo images

Cao,

Derhaag,

Coonen

et al. 2024

Human Reproduction

View full text Add to dashboard Cite

show abstract

“…The series of clinical Turing tests that were performed provided a robust validation method to confirm the realism of the synthetic ECG dataset. Whilst Turing tests have been used to validate other forms of synthetic medical imaging data, [27][28][29][30] our iterative methodology is novel as it integrates healthcare professionals' feedback to enhance the realism of the synthetic images. The methodology described in this study presents a framework that can be used by other disciplines to generate large, life-like synthetic patient datasets reducing the requirement to prospectively create new patient datasets.…”

Section: Discussionmentioning

confidence: 99%

“…A series of visual Turing tests were designed using previously published healthcare research utilising image-based Turing testing [27][28][29][30] and conducted to assess the fidelity of synthetic ECG images via an online survey (Qualtrics, Provo, UT). In all rounds of Turing tests, healthcare professionals were provided with a series of 60 images comprising 30 synthetically created ECGs and 30 photographs of real-world ECGs.…”

Section: Clinical Turing Testsmentioning

confidence: 99%

GenECG: A synthetic image-based ECG dataset to augment artificial intelligence-enhanced algorithm development

Bodagh,

Tun,

Barton

et al. 2023

Preprint

View full text Add to dashboard Cite

Artificial intelligence-enhanced electrocardiogram (AI-ECG) analysis has the potential to transform care of cardiovascular disease patients. Most algorithms rely on digitised signal data and are unable to analyse paper-based ECGs, which remain in use in numerous clinical settings. An image-based ECG dataset incorporating artefacts common to paper-based ECGs, which are typically scanned or photographed into electronic health records, could facilitate development of clinically useful image-based algorithms. This paper describes the creation of GenECG, a high-fidelity, synthetic image-based dataset containing 21,799 ECGs with artefacts encountered in routine care. Iterative clinical Turing tests confirmed the realism of the synthetic ECGs: expert observer accuracy of discrimination between real-world and synthetic ECGs fell from 63.9% (95% CI 58.0%- 69.8%) to 53.3% (95% CI: 48.6%-58.1%) over three rounds of testing, indicating that observers could not distinguish between synthetic and real ECGs. GenECG is the first publicly available synthetic image-based ECG dataset to pass a clinical Turing test. The dataset will enable image-based AI-ECG algorithm development, ensuring the translation of AI-ECG research developments to the clinical workspace.

show abstract

“…Synthetic Datasets Goodfellow et al ( 2014) proposed GAN as a new generative modeling framework [14] to synthesize new data with the same characteristics from training examples, visually approximating the training data set. Various GANbased methods have been proposed for image synthesis in recent years [15], [16], [17], [18], [19], [20], [21], [22], [23], and [24] with applications spreading rapidly from computer vision and machine learning communities to domain-specific areas such as medical [25] [26], [27], [28], [29], and remote sensing [30], [31], [32] [33], [34], [35], [36], [37], [38], [39], [40], and [41]; industrial process [42], [43], [44], [45], [46], [47], and [48]; and agriculture [49], [50], [51], [52].…”

Section: B Gan (Generative Adversarial Network) To Producementioning

confidence: 99%

Oil Palm USB (Unstripped Bunch) Detector Trained on Synthetic Images Generated by PGGAN

Aji,

Bin Ghazali,

Akbar

2023

Journal of Robotics and Control

View full text Add to dashboard Cite

Identifying Unstriped Bunches (USB) is a pivotal challenge in palm oil production, contributing to reduced mill efficiency. Existing manual detection methods are proven time-consuming and prone to inaccuracies. Therefore, we propose an innovative solution harnessing computer vision technology. Specifically, we leverage the Faster R-CNN (Region-based Convolution Neural Network), a robust object detection algorithm, and complement it with Progressive Growing Generative Adversarial Networks (PGGAN) for synthetic image generation. Nevertheless, a scarcity of authentic USB images may hinder the application of Faster R-CNN. Herein, PGGAN is assumed to be pivotal in generating synthetic images of Empty Fruit Bunches (EFB) and USB. Our approach pairs synthetic images with authentic ones to train the Faster R-CNN. The VGG16 feature generator serves as the architectural backbone, fostering enhanced learning. According to our experimental results, USB detectors that were trained solely with authentic images resulted in an accuracy of 77.1%, which highlights the potential of this methodology. However, employing solely synthetic images leads to a slightly reduced accuracy of 75.3%. Strikingly, the fusion of authentic and synthetic images in a balanced ratio of 1:1 fuels a remarkable accuracy surge to 87.9%, signifying a 10.1% improvement. This innovative amalgamation underscores the potential of synthetic data augmentation in refining detection systems. By amalgamating authentic and synthetic data, we unlock a novel dimension of accuracy in USB detection, which was previously unattainable. This contribution holds significant implications for the industry, ensuring further exploration into advanced data synthesis techniques and refining detection models.

show abstract

Image Turing test and its applications on synthetic chest radiographs by using the progressive growing generative adversarial network

Cited by 5 publications

References 37 publications

Generative artificial intelligence to produce high-fidelity blastocyst-stage embryo images

Generative artificial intelligence to produce high-fidelity blastocyst-stage embryo images

GenECG: A synthetic image-based ECG dataset to augment artificial intelligence-enhanced algorithm development

Oil Palm USB (Unstripped Bunch) Detector Trained on Synthetic Images Generated by PGGAN

Contact Info

Product

Resources

About