Diagnosis of skin diseases in the era of deep learning and mobile technology

Göçeri, Evgin

doi:10.1016/j.compbiomed.2021.104458

Cited by 89 publications

(29 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The number of training samples mainly depends on the targets' motion cycle and the camera's sampling frequency. Usually, manual operation is used to collect training samples, which is less time-effective given the large-scale data requirement of NN training [ 26 – 28 ]. Against such a dilemma, the proposed VML method employs an automated sample data generation process.…”

Section: Virtual Vst Based On Lightweight Deep Learning Modelmentioning

confidence: 99%

Applying Lightweight Deep Learning-Based Virtual Vision Sensing Technology to Realize and Develop New Media Interactive Art Installation

Luo

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

The work intends to optimize the situation that interactive art devices and remote control based on traditional technology cannot meet people’s actual needs to a certain extent. With the assistance of Lightweight Deep Learning (LDL) models, Interactive Artistic Installation (IAI) shows excellent creative potential in terms of dimension, space, and sense. Virtual Vision Sensing Technology (VST) explores the emotional semantics in the human-machine environment with the help of interactive art, finds the emotional interaction elements between human and machine, and promotes Human-Computer Interaction (HCI). From the perspective of the media elements of interactive art, this paper reviews the virtual VST that subverts the expression of interactive art. Then, from the perspective of artistic creation, the impact of virtual VST on IAI thinking, methods, and artistic experience is analyzed. Thereupon, a scene construction method is designed where the physical equipment is premodeled. The model is loaded in real time with visual information. The proposed method does not require complex vision and laser scanning equipment or high-configured computer systems. The proposed new media IAI model realizes the real-time loading of the scene model. According to the physical equipment dynamic information obtained by the visual data acquisition system, the proposed method can keep the virtual scene and physical models in motion synchronization. Finally, experiment results corroborate that the environment will significantly interfere with the experimental results. The training data set with boundary occlusion will be more suitable for model training and better test results (about 97% accuracy). Hence, the research content can make the Virtual Reality works have better performance, especially the sense of experience from the perspective of aesthetics. Meanwhile, it also enriches the research theory in the field of new media art installation technology.

show abstract

Section: Virtual Vst Based On Lightweight Deep Learning Modelmentioning

confidence: 99%

Applying Lightweight Deep Learning-Based Virtual Vision Sensing Technology to Realize and Develop New Media Interactive Art Installation

Luo

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“… 2017 ; Goceri 2016 ; Göçeri et al. 2015 ; Göçeri 2013 ), and deep learning-based approaches have been widely used recently (Goceri 2021 ; Göçeri 2020 ). The evolving medical imaging acquisition system (Alexander et al.…”

Section: Introductionmentioning

confidence: 99%

Modality specific U-Net variants for biomedical image segmentation: a survey

2022

View full text Add to dashboard Cite

With the advent of advancements in deep learning approaches, such as deep convolution neural network, residual neural network, adversarial network; U-Net architectures are most widely utilized in biomedical image segmentation to address the automation in identification and detection of the target regions or sub-regions. In recent studies, U-Net based approaches have illustrated state-of-the-art performance in different applications for the development of computer-aided diagnosis systems for early diagnosis and treatment of diseases such as brain tumor, lung cancer, alzheimer, breast cancer, etc., using various modalities. This article contributes in presenting the success of these approaches by describing the U-Net framework, followed by the comprehensive analysis of the U-Net variants by performing (1) inter-modality, and (2) intra-modality categorization to establish better insights into the associated challenges and solutions. Besides, this article also highlights the contribution of U-Net based frameworks in the ongoing pandemic, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as COVID-19. Finally, the strengths and similarities of these U-Net variants are analysed along with the challenges involved in biomedical image segmentation to uncover promising future research directions in this area.

show abstract

“…Overall Losses. Although hybrid loss functions have been applied for different purposes in recent works [12,13], we adopt the multi-task losses as:…”

Section: Frameworkmentioning

confidence: 99%

UniInst: Unique Representation for End-to-End Instance Segmentation

Ou¹,

Yang²,

Ma³

et al. 2022

Preprint

View full text Add to dashboard Cite

Existing instance segmentation methods have achieved impressive performance but still suffer from a common dilemma: redundant representations (e.g., multiple boxes, grids, and anchor points) are inferred for one instance, which leads to multiple duplicated predictions. Thus, mainstream methods usually rely on a hand-designed non-maximum suppression (NMS) postprocessing step to select the optimal prediction result, which hinders endto-end training. To address this issue, we propose a box-free and NMS-free end-to-end instance segmentation framework, termed UniInst, that yields only one unique representation for each instance. Specifically, we design an instance-aware one-to-one assignment scheme, namely Only Yield One Representation (OYOR), which dynamically assigns one unique representation to each instance according to the matching quality between predictions and ground truths. Then, a novel prediction re-ranking strategy is elegantly integrated into the framework to address the misalignment between the classification score and the mask quality, enabling the learned representation to be more discriminative. With these techniques, our UniInst, the first FCN-based end-to-end instance segmentation framework, achieves competitive performance, e.g., 39.0 mask AP using ResNet-50-FPN and 40.2 mask AP using ResNet-101-FPN, against mainstream methods on COCO test-dev. Moreover, the proposed instance-aware method is robust to occlusion scenes, outperforming common baselines by remarkable mask AP on the * Equal Contribution

show abstract

Diagnosis of skin diseases in the era of deep learning and mobile technology

Cited by 89 publications

References 45 publications

Applying Lightweight Deep Learning-Based Virtual Vision Sensing Technology to Realize and Develop New Media Interactive Art Installation

Applying Lightweight Deep Learning-Based Virtual Vision Sensing Technology to Realize and Develop New Media Interactive Art Installation

Modality specific U-Net variants for biomedical image segmentation: a survey

UniInst: Unique Representation for End-to-End Instance Segmentation

Contact Info

Product

Resources

About