Melanoma Skin Lesion Classification Using Neural Networks: A systematic review

Standard resection of pediatric facial skin lesions consists of lenticular excision and linear closure. This one-stage procedure for circular lesions results in a linear scar 3 times longer than the diameter of the removed specimen. Circular excision and purse-string closure has been described for infantile hemangiomas to reduce the length of scar. The purpose of this study was to analyze the application of this technique for any type of circular facial skin lesion in the pediatric population. Records of consecutive pediatric patients with facial skin lesions treated with circular excision and purse-string closure from 2007-2014 were reviewed. Patient age, sex, type of lesion, location, and size were recorded. Number of stages necessary to remove the area and complications were analyzed. Seventy-seven children (74% female) underwent circular excision and purse-string closure for an infantile hemangioma (46%), pigmented nevus (27%), Spitz nevus (7%), pilomatrixoma (5%), pyogenic granuloma (5%), vascular malformation (4%), or another type of skin lesion (6%). Age at the time of resection was 6.0 years (range 4 months-17 years) and mean lesion area was 3.9 cm (range 0.2-19.6); 30% of patients underwent a second procedure and no infection or wound dehiscence occurred. Circular excision and purse-string closure is an effective technique to manage any type of circular skin lesion in the pediatric population. It is particularly useful for lesions on the face because it limits the length of a scar. A subset of patients may benefit from second procedure to convert the circular scar from a circle into a line.

Skin cancer is among the life-threatening cancers, but unlike most cancers, skin cancer is observable and can be detected in early stages, yet not many are aware of its detectability. There are mainly three types of skin cancers, which are basal cell carcinoma, squamous cell carcinoma, and melanoma, where melanoma is the most dangerous type of cancer with a very low survival rate. Skin cancers are not painful, most of the time, even though they appear to be visibly distressing it makes them easily detectable, as cancer is nothing but the abnormal growth of skin cells. A person can detect if a skin lesion is cancerous by taking a picture. Deep neural networks can be used to classify the type of cancer. This can be done by collecting and feeding several clinical images of cancerous skin lesions, segmentation, removing noise, etc., to a deep neural network to train on before detecting cancerous lesions. Our data was scraped from the Internet and few images were collected from the HAM10000 dataset, ISIC Archive, and scraped images from the Web. Every class has 3552 images which are a total of 10,656 images; image augmentation was used to generate images to make all classes have an equal number of images. The first model was a basic CNN model that trained, several times changing the hyperparameter values to fine-tune the model to give accurate results, which gave us 86.5% accuracy and implemented transfer learning with the ImageNet weights of different ImageNet models, where ResNet50 gave us the highest accuracy of 95.6%. We have deployed this into a Web application using JavaScript and tensorflow.js.

Diagnosis of skin lesions is a very challenging task due to high inter-class similarities and intra-class variations between lesions in terms of color, size, site and appearance. As a result, as is the case with many types of cancer, early detection of skin cancer is vital for survival. Advances in artificial intelligence, in particular, deep learning have enabled to design and implementation of intelligence-based lesion detection and classification systems that are based on visible light images that are capable of performing early and accurate diagnosis of different skin diseases. In most cases, the precision of these methods has reached a level of accuracy that is comparable to that achieved by a qualified dermatologist. This work presents potential skin lesion classification solutions based on the datasets taken from the most recent publicly available “Skin Lesion Analysis Towards Melanoma Detection” grand challenges ISIC 2018. The proposed classification approach leverages convolutional neural networks (CNN) and transfer learning to enhance skin classification. Different pre-trained models were applied, including VGG-Net, ResNet50, InceptionV3, Xception and DenseNet121. Additionally, the heavy class imbalance is examined as a critical problem for this dataset and multiple balancing techniques, such as weight balancing and data augmentation, are considered. Finally, an ensemble approach is evaluated by combining and averaging several CNN architectures to classify the seven different types of skin lesion. The experimental results indicate that the proposed frameworks exhibit promising results when compared with ISIC 2018 challenge live leaderboard.

Monkeypox (MPX) is a zoonotic disease caused by an orthopoxvirus usually causing self-limiting outbreaks in Africa. Since May 2022, an unprecedented outbreak has spread globally.1 The classical presentation of MPX disease associates fever, headache, myalgia, lymphadenopathy, sore throat and pleiomorphic skin lesions. Typical skin lesions are papules, vesicles and pustules that ulcerate, become umbilicated and then crops. However, the clinical presentation of MPX disease is sometimes aspecific and can mislead diagnosis.2, 3 We conducted a single centre retrospective study in Paris, France, to analyse MPX patients presenting with an erythematous maculopapular rash. All patients gave oral consent to participate and the study was approved by a Research Ethics Committee. Between 21st May and 28th July 2022, 439 patients had a PCR-confirmed MPX infection in this centre. Thirty patients (30/439, 6.8%) presented an erythematous maculopapular rash (Table 1; Figure 1). Patients were all men except one male to female transgender and 93% had sex with men. Among the 30 patients, 25 (83%) presented concomitant typical MPX vesiculopustular skin lesions and 15 (50%) presented concomitant pharyngitis (Figure 1). Of note, five patients (17%) had no typical MPX skin lesions and presented only with an erythematous maculopapular rash and a pharyngitis. Rashes appeared in a median of 3.5 days [IQR 2–6] after symptoms' onset. Dermatological examination revealed aspecific maculopapular rashes, always affecting the trunk and sometimes extending to the limbs or face. Three patients were hospitalized: one for aphagious pharyngitis, one for blepharitis and one for phimosis with urinary retention. All rashes resolved in a few days without specific treatment. The prescription of an antibiotic therapy before the apparition of the rash was reported in 17 cases (59%), mostly amoxicillin (n = 13) for suspected bacterial pharyngitis. In patients with antibiotics, the rash followed antibiotic use after a median of 1 day [IQR 1–2]. Among the 13 patients for whom EBV blood PCR was done, all had a replication. Of note, EBV blood PCR was performed in seven patients of the centre with MPX without a rash and found EBV replication in the seven cases. Histopathological findings were similar in the six patients who had a cutaneous biopsy to rule out a severe drug reaction. It consisted of a superficial lymphocytic infiltrate consistent with maculopapular exanthema of drug or viral origin. The immunohistochemical EBV marker (EBER) was always negative and there was no specific MPX histological pattern to guide diagnosis. In this centre, rashes frequency is estimated at 6.8% of MPX patients, which makes it less frequent than other recent reviews (8.0%–13.7%).3-5 Half the patients in this study had an association of rash and pharyngitis leading to antibiotics prescription to treat a suspected bacterial pharyngitis. MPX diagnosis may reduce the use of unnecessary antibiotic prescribing. MPX rashes could be related to different physiopathological mechanisms including an aspecific MPX cutaneous tropism or concomitant viral EBV replication with cutaneous manifestation.6, 7 Rashes onset in 59% of patients after antibiotic therapy might also suggest a drug reaction or an infectious mononucleosis-like antibiotic reaction.8 A primary EBV infection is much less probable even though clinical presentation can be misleading.9 An erythematous maculopapular rash was observed in 13 patients without prior antibiotic therapy making it a specific skin manifestation of MPX infection. Maculopapular rashes are not uncommon in MPX. They are associated to pharyngitis in half of the cases and to typical skin lesions in 83% of the cases. In an epidemic context, MPX infection should thus be considered in patients with an erythematous maculopapular rash even without associated skin lesions. None. There are no conflict of interest. The patients in this article have given written informed consent to publication of their case details.

Multi-class segmentation of skin lesions via joint dictionary learning

A novel nonlinear automated multi-class skin lesion detection system using soft-attention based convolutional neural networks

Melanoma is a type of skin lesion that is less common than other types of skin lesions, but it is fast growing and spreading. Therefore, it is classified as a serious disease that directly threatens human health and life. Recently, the number of deaths due to this disease has increased significantly. Thus, researchers are interested in creating computer-aided diagnostic systems that aid in the proper diagnosis and detection of these lesions from dermoscopy images. Relying on manual diagnosis is time consuming in addition to requiring enough experience from dermatologists. Current skin lesion segmentation systems use deep convolutional neural networks to detect skin lesions from RGB dermoscopy images. However, relying on RGB color model is not always the optimal choice to train such networks because some fine details of lesion parts in the dermoscopy images can not clearly appear using RGB color model. Other color models exhibit invariant features of the dermoscopy images so that they can improve the performance of deep neural networks. In the proposed Color Invariant U-Net (CIU-Net) model, a color mixture block is added at the beginning of the contracting path of U-Net. The color mixture block acts as a mixer to learn the fusion of various input color models and create a new one with three channels. Furthermore, a new channel-attention module is included in the connection path between encoder and decoder paths. This channel attention module is developed to enrich the extracted color features. From the experimental result, we found that the proposed CIU-Net works in harmony with the new proposed hybrid loss function to enhance skin segmentation results. The performance of the proposed CIU-Net architecture is evaluated using ISIC 2018 dataset and the results are compared with other recent approaches. Our proposed method outperformed other recent approaches and achieved the best Dice and Jaccard coefficient with values 92.56% and 91.40%, respectively.

Objective To compare the diagnostic accuracies of deep convolutional neural network (CNN) and dermatologists for pigmented nevus and seborrheic keratosis. Methods CNN network ResNet-50 was trained with 5 094 dermoscopic images of pigmented nevus and seborrheic keratosis using transfer learning, so as to establish a CNN two-classification model. Then, this model was applied to the automatic classification of 30 dermoscopic images of pigmented nevus and 30 dermoscopic images of seborrheic keratosis. Meanwhile, in combination with clinical photos of skin lesions, 95 experienced dermatologists who had received dermoscopy training gave their diagnosis for the above 60 dermoscopic images. The diagnostic accuracies were compared between the two methods, and misclassified images were further analyzed. Results The CNN automatic classification model had the diagnostic accuracies of 100% (30/30) and 76.67% (23/30) for pigmented nevus and seborrheic keratosis respectively, and the total accuracy was 88.33% (53/60) . The average diagnostic accuracies of 95 dermatologists were 82.98% (25.8/30) and 85.96% (24.9/30) for pigmented nevus and seborrheic keratosis respectively, and the total accuracy was 84.47% (50.7/60) . There were no significant differences in the diagnostic accuracies for pigmented nevus or seborrheic keratosis between the CNN automatic classification model and 95 dermatologists (χ2 = 0.38, P > 0.05) . The dermoscopic images misclassified by CNN were divided into 3 categories: special - type lesions with high pigment content and marked keratosis, typical skin lesions with interference factors, and typical skin lesions without definite reasons for misclassification. Conclusions The performance of CNN automatic classification model is similar to that of experienced dermatologists in the two classification of pigmented nevus and seborrheic keratosis. The reasons for misclassification by CNN still need to be explored by dermatologists and professionals in artificial intelligence. Key words: Nevus, pigmented; Keratosis, seborrheic; Dermoscopy; Neural networks (computer); Artificial intelligence; Convolutional neural network

The paper proposes a system for determining malignant skin neoplasms. The use of convolutional neural networks for determining skin tumors from images is considered. A convolutional neural network of deep learning has been developed and modeled, which allows you to determine and classify pigmented skin lesions by examining photographs. The article proposes a system for determining malignant skin neoplasms. The proposed neural network has the basic parameters of the VGG-A architecture with a maximum number of epoch training of 10. The accuracy of the determination of the proposed model of the convolutional neural network is at least 77%. The minimum learning loss function was 0.5577. As a result of the work, a database of training photos of real pigmented skin formations taken from the international open archive ISIC Melanoma Project was used, which is a database of digital images of various types of skin lesions, was formed. Using the proposed model can be of great help in determining and diagnosing malignant skin lesions by dermatologists.

SummaryMelanoma is a type of a skin cancer or lesion which has the detrimental ramifications on the human health but with early diagnosis it can be cured easily. The actual identification of skin lesion is very challenging because of factors like a very minute difference between lesion and skin and it is very difficult to differentiate among skin cancer types due to visual comparability. Hence an autonomous system for the diagnosis of true skin cancer type is very useful. In this article, we took the leverage of ensemble learning by combining the features of deep learning architectures with traditional features extraction approaches. For segmentation, we have two pipelines for the feature extraction. We extract the features through traditional split and merge approach as well as from deep learning algorithms of contextual encoding along with the attention mechanism. Later we combine the features of both architectures and predict the segmented region through intersection over union mechanism. After that segmented region is classified into three types of skin lesion using hybrid features of Alex‐Net and VGG‐16 through the transfer learning approach. The evaluation has been performed using the ISIC and PH2 datasets for which achieved segmentation accuracy is 97.8% and 96.7%, respectively. Moreover, hybrid classification network able to attain the 98.2% accuracy.

Skin cancer is one of the most fatal skin lesions, capable of leading to fatality if not detected in its early stages. The characteristics of skin lesions are similar in many of the early stages of skin lesions. The AI in categorizing diverse types of skin lesions significantly contributes to and helps dermatologists to preserve patients' lives. This study introduces a novel approach that capitalizes on the strengths of hybrid systems of Convolutional Neural Network (CNN) models to extract intricate features from dermoscopy images with Random Forest (Rf) and Feed Forward Neural Networks (FFNN) networks, leading to the development of hybrid systems that have superior capabilities early detection of all types of skin lesions. By integrating multiple CNN features, the proposed methods aim to improve the robustness and discriminatory capabilities of the AI system. The dermoscopy images were optimized for the ISIC2019 dataset. Then, the area of the lesions was segmented and isolated from the rest of the image by a Gradient Vector Flow (GVF) algorithm. The first strategy for dermoscopy image analysis for early diagnosis of skin lesions is by the CNN-RF and CNN-FFNN hybrid models. CNN models (DenseNet121, MobileNet, and VGG19) receive a region of interest (skin lesions) and produce highly representative feature maps for each lesion. The second strategy to analyze the area of skin lesions and diagnose their type by means of CNN-RF and CNN-FFNN hybrid models based on the features of the combined CNN models. Hybrid models based on combined CNN features have achieved promising results for diagnosing dermoscopy images of the ISIC 2019 dataset and distinguishing skin cancers from other skin lesions. The Dense-Net121-MobileNet-RF hybrid model achieved an AUC of 95.7%, an accuracy of 97.7%, a precision of 93.65%, a sensitivity of 91.93%, and a specificity of 99.49%.

Skin cancer is widely menacing forms of cancer in North America and South East Asia and some part of Australia also. The main reason of skin cancer is caused by damaged deoxyribonucleic acid (DNA) in skin cells of human body which is inherited from genetic disorder or mutations on the skins. Skin cancer is to gradually spreading over other body parts with acute pain and it is only curable in initial stages. That is always recommended to detect at early stages of cancer as we know that there are four stages of cancer. The skin cancer has high mortality rate all over the globe as compare to other types of cancer and its treatments are very expensive. This paper presents a detailed review of deep learning techniques like convolutional neural network for the early detection of skin cancer and their types. We have develop a 2-D CNN model and evaluated the model with different parameters and finally evaluated model with data augmentation and predict the incorrect probability for different types of Skin cancer for HAM 1000 datasets.KeywordsSkin diseasesDeep learningDermoscopic imagesConvolutional neural networks (CNN)RegularizationData augmentation

Impact of Cutaneous Involvement on the Clinical Outcome of Adult T-Cell Leukemia/Lymphoma: A Study from the Latin American Group of Lymphoproliferative Disorders (GELL)

Multiple skin lesions diagnostics via integrated deep convolutional networks for segmentation and classification

In recent years, the position location applications have increasingly. In this paper, we will use multiple Back-Propagation neural networks with genetic algorithm GA for a radio frequency identification RFID indoor location system to provide location services named indoor location with multiple neural networks and genetic algorithms ILMNGA. In Section 1, we collect received signal strength RSS information from reference points to train the neural network models. In Section 2, genetic algorithm GA is used to find the weight of each neural network based on the performance of each neural network. Finally, we input the RSS information of each tracking object into the model that will provide the location of tracking objects based on the RSS information. The location will be integrated using the weights produced by the GA. The experiment conducted our methodology can provide better accuracy than a single neural network.