Data Collection
We collect high-quality data across multiple data types, including text, speech & audio, image, and video, in a variety of scenarios.
Text
- E-commerce data
- Music titles
- Handwritten notes
- Abbreviations and acronyms
- Geographical locations
- Ads
- Textbooks
- News
- Social media
- Legal documents
Speech & Audio
- Speech-to-Text (STT)
- Speech output by phones/embedded devices
- Single/multi-person speech
- Near/far field speech
- Indoor/outdoor speech
- Speech at recording studio/public places
- Reading/speaking/conversation
Image
- Multi-category images
- Multi-scene images
- Faces & expressions
- Graphic symbols
- Home construction
- Traffic elements
- Location-based images
Video
- Videos of specified persons and actions
- Videos in specified environments
- Aerial photography
Data Annotation
Our data annotation service accurately annotates, classifies, and analyzes unstructured multilingual data needed to train machine learning applications to ensure accurate results: cleaning, extraction, and annotation at scale for text, speech, audio, image, video, and other special data.
Text
- Text transcription into phonemic or phonetic equivalent
- Semantic annotation
- Emotional tone annotation
- Parallel corpora annotation
- Text keywords
- Content classification
- Sentence structure
- Semantics, stem and lexicon
Speech & Audio
- Speech transcription
- Speech screening
- Background environment noise
- Speaker role
- TTS accent, intonation and rhythm
Image
- Image classification
- Image tag classification
- Image relevance evaluation
- Image content extraction
- Invalid image filtering
- Image text extraction
- Facial expression annotation
Video
- Video segmentation
- Video transcription
- Behavioral trajectory
- Human expression & emotion
- Video quality screening
- Content moderation
- Aerial photography
Data Evaluation
ECI’s data evaluation service is designed to help you improve the quality and accuracy of online information, optimize content moderation, and ensure search engine relevance.
- Optimize content moderation
We protect your brand by assisting you in ensuring the comments, reviews and feedback left by users with Chatbots and other social media channels are free of hate speech or other malicious content. - Chatbot
If your Chatbot is built on mature data, we can help it create chat modes and trigger instant answers to frequently asked questions to respond to customers more quickly. - Social media
Want to know what your end users are saying about you on social networks? We can help you ensure that your users get the best experience and maintain loyalty to your brand. - Search relevance evaluation
Our customized solutions allow you to improve overall search relevance. Our services include the evaluation of specific search results and even the entire search engine results page.
Annotation Methods
Polyline Annotation
This is usually used to annotate lanes for self-driving vehicles to define the positions of linear objects precisely without excessive noise and white spaces.
Rectangle Annotation
This is the simplest image annotation type. A large volume of rectangle annotation data can train the model to recognize the objects you need.
Key-Point Annotation
It is used to track the movement trajectories of key points for statistical models (such as foot traffic calculation models for commercial sites) as well as gesture or facial recognition models.
Polygon Annotation
This annotation features higher precision to avoid visual model deviation caused by a large amount of white space and extra noise. Common use cases include robotic grabber, medical image recognition, satellite image recognition, etc.
Semantic Segmentation
As the most precise annotation method in image annotation, it involves labeling each pixel in an image, allowing you to make full use of limited data sources.
Image Classification
This is a very important data cleaning task, especially for image recognition-based applications. It enables you to classify an image according to its visual content, paving a way for more complex annotation tasks in later stages.
Industries We Serve
Intelligent Healthcare
- Image Annotation
- Drug Discovery
- Prescription Transcription
Intelligent transportation
- Traffic Object Detection/Positioning
- Panoramic Semantic Segmentation of Traffic Images
- Trajectory Tracking Video Annotation
Retail
- E-commerce
- SKU Recognition
- Unmanned Supermarket
- Visual Search
Intelligent Home
- Vehicle-Related Data
- Pedestrian-Related Data
- Road Video Surveillance Data
Intelligent Security
- Vehicle-Related Data
- Pedestrian-Related Data
- Road Video Surveillance Data
- Public Place Video Surveillance Data
