- 10 Oct 2023
- 3 Minutes to read
- Updated on 10 Oct 2023
- 3 Minutes to read
How many images do I need to upload for model training?
Since every use case is different, there is no one number that works in every situation. This number generally depends on these factors:
- How elaborate the object to identify is
- How many types of objects there are
- How the object looks
For example, a hammer with a broken handle would require fewer images than a printed circuit board that has several intricate parts.
Tips for Object Detection Projects
Models in Object Detection projects learn based on the contents of the bounding boxes (the areas you identify as the Ground Truth). Therefore, it's important to get a well-represented spread of bounding boxes for each class.
For complex projects (like circuit boards), we recommend you include a minimum of 50 bounding boxes per class. If one image has five bounding boxes, then this counts as five, not one.
Tips for Classification Projects
Each image can only have one class applied to it. Therefore, if an image meets the criteria for more than one class, we recommend that you don't include it in training because it could confuse the model.
For complex projects (like circuit boards), we recommend you include a minimum of 50 images per class.
Tips for Segmentation Projects
Models in Segmentation projects learn based on the specific pixels marked by the lines or shapes you add to the images (the areas you identify as the Ground Truth).
Therefore, it's important to get a well-represented spread of marked areas for each class.
Typically, a Segmentation project requires fewer images than an Object Detection project. This is because a Segmentation project allows you to label areas more precisely.
For complex projects (like circuit boards), we recommend you include about 20 to 30 labeled areas per class. If one image has five labeled areas, then this counts as five, not one.
Tips for Visual Prompting Projects
You only need to label a few small areas of images to train a Visual Prompting model. For more information, go to Visual Prompting.
What image file types does LandingLens support?
You can upload the following file types to LandingLens: PNG, BMP, JPG, JPEG, and MPO.
Object Detection also supports Pascal VOC files. Pascal VOC (Visual Object Classes) is a file format that includes the label details of its paired image. It essentially tells LandingLens where a label is on the associated image. For more information, go to Upload Labeled Images to Object Detection Projects.
What types of labeling tools are available?
LandingLens offers different project types, and each type has different labeling tools.
|Use bounding boxes (rectangles) to outline objects.
|Classify images after upload using classes. Or, upload "classified" folders; LandingLens will automatically classify images based on the titles of the folders. (Classification looks at all the pixels in an image.)
|Segmentation has multiple labeling tools:
|Use a Brush to "paint" specific pixels.
|Separately upload "normal" and "abnormal" images. Similar to Classification, Anomaly Detection considers all pixels in an image.
Anomaly Detection is only available to legacy "classic" workflow users.
Do you have tools to evaluate consistency among different Labelers?
Yes! LandingLens offers Agreement-Based Labeling that allows multiple Labelers to label the same images. LandingLens then ranks those images based on levels of consistency.
What types of augmentation tools are available?
- Random brightness
- Blur, motion blur, and gaussian blur
- Hue saturation
- Random contrast
- Horizontal and vertical flips
- Random rotate
Can I upload pre-labeled images to LandingLens?
Can I upload zipped files to LandingLens?
All files must be unzipped before uploading them to LandingLens.
Does LandingLens keep folder hierarchies on upload?
LandingLens simplifies folder structures and flattens all files. For more information, go to Upload Folders.