According to the relevant person in charge of the appraisal agency, ChinaStarsAutograph has collected original handwriting samples signed by more than 2,000 celebrities around the world, and now it can provide services such as artist's original handwriting appraisal, normal handwriting appraisal and signature handwriting appraisal, and can provide relevant certificates of appraisal results.

CSA staff conduct professional appraisal through the world's advanced Lecun handwriting appraisal technology, judge the signature to be appraised and make appraisal results. At the same time, in order to improve the accuracy of signature identification, CSA institutions will regularly update celebrity handwriting templates and regularly train appraisers.

In addition, the organization also has several special commissioners who often participate in the artist's recent activities in order to provide photos of the artist's live activities. At present, CSA has opened online official website and the latest certification. For details of CSA, fans and friends can inquire through official website if they have any questions.

Any artist's signature recognized by fans in CSA company can also be checked for authenticity through the inquiry code that comes with the certificate, thus ensuring the authenticity of the signed products purchased.

Identification technology:

The agency will send staff to the star's recent activities to sign and sign in, and will also arrange photos with the star as appropriate, and will provide on-site work permits, live videos or photos as evidence.

In fact, character recognition has always been a major application direction of deep learning. Deep learning models such as CNN/LSTM have a long history in the application of character recognition. In the 1990s, pioneers of deep learning, such as Y.Lecun, used neural networks to solve character recognition very early.

1998, Lecun and Bengio cooperated to design LeNet5 to solve the handwritten numeral recognition problem, which was their demonstration in Bell Laboratories. Later, scholars including Microsoft Cambridge Research Institute achieved a very low error rate (0.4%) with CNN in 2003. CSA introduced this recognition technology to ensure the accuracy of handwriting.