Recent reports on dermoscopy techniques for skin cancer diagnosis.

S. no.	Description	Upside	Downside	Acc, Se, Sp	Ref.
1	Dermoscopy significantly has higher discriminating power than the clinical analysis. The sensitivity and the specificity ranges obtained from diagnosis of melanoma were found to be 0.75 to 0.96 and 0.79 to 0.98 respectively	More accurate than clinical examination for the diagnosis of melanoma in a pigmented skin lesion	Requires experience for better diagnosis	Se: 0.75–0.96 and Sp: 0.79–0.9	54
2	Compared the dermatoscopic features of lento senilis and lentigo malinga on the face by using logistic regression analysis	Analysis is easy	Resolution is low	Se: 93.8% and Sp: 52.3%	55
3	Dermoscopic images of the skin lesions were analyzed using 2-step diagnostic methods	Good computational capability	Various algorithms are needed	Se: 64.8% and Sp: 72.8%	56
4	Dermoscopic images of 20 pigmented skin lesions were evaluated based upon the menzie's method and ABCD rule and pattern analysis. It was found that results of pattern analysis were comparatively more accurate than any other methods	Web-based training is an effective tool for teaching dermoscopy	Involves training of practioners	Acc: 62.8%	57
5	The vascular structure of melanocytic and non-melanocytic skin tumours were evaluated based upon the morphological features	High resolution	Analysis of distinctive vascular structures is required	Se: 81.1%	58
6	Dermoscopy helps in early diagnosis of melanoma cancer by in vivo methods	Early diagnosis of melanoma	Resolution is low	Acc: 5% to 30%	59
7	Primary physicians made study on 2522 skin cancer subjects and the accuracy level were compared with the existing system	Improves the ability of physicians to triage lesions suggestive of skin cancer	Involves physician training, algorithm and expert consultation	Se: 54.1%, Sp: 71.3%	60
8	The studies were made on 35 healthy pregnant women and 35 age-matched female controls. The analysis showed that the pregnancy leads to significant modifications in PSL, especially with respect to globules, pigment network, and architectural order or disorder	Local intensity variant is done	Consumes more time to diagnosis the cancer	Se: 79.3%, Sp: 93.18%	66
9	In this method, the dermoscopic images obtained from the subject were analyzed based on border detection	Fast and accurate border detection	Requires a series of process algorithms	—	62
10	Through the dermosccpy technique the non-melanoma skin cancer were detected and the accuracy was high	Identification of skin lesion is clear	Only detects the non-melanoma cancer.	—	68
11	In dermoscopic images, the diagnosis of melanoma can be easily identified by irregular streaks	Accuracy level is high	Only includes the analysis of irregular streaks	Acc: 76.1%	69
12	Demonstrated the classification of skin lesions using a single deep convolutional neural network (CNN), trained end-to-end from images directly, using only pixels and disease labels as inputs	The CNN achieves performance on par with all tested experts	Involves CNN computing platform	Acc: 72.1 ± 0.9%	70