Cimation (M) Sdn Bhd

Mech-DLK

Mech-DLK enables users to easily train deep learning models for various parts. High-precision deep learning algorithms guarantee extremely high accuracy with fewer parameters. Advanced data augmentation enables users to train models with smaller image sets.

Efficiently Recognize

With object detection, image classification, instance segmentation, and semantic segmentation, these allow users to train models faster and easily solve the most demanding applications like overlapping object recognition and classification, high-precision measurement, etc.

Ease to Use

Mech-DLK enables users to train deep-learning models for a variety of parts with ease. Requiring smaller sample images sets, the high-precision deep learning algorithms ensure high accuracy while keeping parameters tweaking to a minimum. Users can skip the hassle of training a model from scratch and just start optimising existing models for their purposes.

Visualized Validation

Mech-DLK can perform model validation, visually display validation results, and show the comparison between validation results and the annotation, which significantly improve the validation efficiency of vision solutions.