information community file
All customer cases - Huazhi Biology
Huazhi Biology
Huazhi Rice Biotechnology Co., Ltd. is a high-tech innovative enterprise in the seed industry with rich experience in R&D and management of multinational seed industry companies. It is committed to building a world-class high-tech biotechnology R&D center and technical service center in the seed industry.
Use product
Support and communication
Speech recognition helps to collect agricultural big data efficiently
Value achievements
The Huazhi Breeding Housekeeper APP developed by Huazhi Rice Biotechnology Co., Ltd. enables front-line breeding information collectors to collect crop names, traits, indicators and other data by voice through mobile phone APP by combining the model training ability and long speech recognition technology of Baidu AI voice self training platform, It has realized the ability of data management and professional data analysis of breeding information entered into the system.

1. The way from handwritten recording to voice acquisition was changed, and the collection of field property information for a test was reduced from 300 man hours to 160 man hours, and the work efficiency was doubled.

2. After scientific training on Baidu's voice self training platform technology, the overall speech recognition rate has increased from 71.09% to 98.75%, effectively solving the problem of low recognition rate of professional words in the industry, and providing more intelligent agricultural companies with integrated solutions.
Case Story
Core demands
The breeding information collector of Huazhi Rice Biotechnology Co., Ltd. needs to carry tools for field observation and handwritten input when collecting information in the field. This method is extremely troublesome and inefficient, so he came up with the idea of optimizing operations by voice input information. After investigating the speech recognition capabilities of many companies in the market, it was found that due to the professionalism of the agricultural industry, the professional terms of plant traits (such as heading date, seedling salt tolerance, leaf sheath color, etc.) could not recognize the correct characters, which delayed the project's landing. Therefore, Huazhi Biology urgently needs to find a voice recognition company with high accuracy of industry vocabulary to help it solve this problem.
Solution
In the subsequent market research, Huazhi Biology learned about Baidu AI voice technology, which can help it quickly realize the zero code self-help training language model under the agricultural semantic environment and the ability to accurately identify long voice content. Now we choose R&D cooperation. When hundreds of text sentences are sorted out from the fixed combination of "experimental field number+plant traits+indicators" through the voice self training platform, they are uploaded and trained. After continuous iterative training, the final model accuracy rate has the ability to identify and use with high accuracy.

At the same time, on the basis of the original system of the Huazhi Breeding Housekeeper APP developed by Huazhi Biology, through embedding the voice acquisition function, it has realized the ability of structural recognition after saying the "experimental field number+plant traits+indicators". Thus, the collection of character data can be realized, and the operation can be carried out at any time. It can get rid of the limitation of relying on the screen for field data collection, and realize agricultural informatization and intelligence.

The acquisition process of voice input information of breeding personnel is as follows:

Step 1: Open Huazhi Breeding Housekeeper APP and click the bottom voice function button;

Step 2: Click voice input on the information entry page, and then directly follow the "test field number+plant traits+indicators";

Step 3: You can see the result of voice input and conversion into text, and you can also view the recording or upload data directly.
Technical capability
Voice technology
Character recognition
Face and Human Body
Image technology
Language and knowledge
video technique