Shutter Stream Automated Background Removal Tool
Shutter Stream Automated Background Removal Tool
There has been an increasing emphasis on product photography with white backgrounds in online retail and website imaging so we decided to explore ways that would help users streamline this process. The Automated Background Removal Tool in the Shutter Stream and Shutter Stream 360 Product Photography software was designed to save users time and effort when removing backgrounds from product images. We explored many different alternatives to automate the background removal process and decided developing image recognition algorithms would prove to be the best methodology as it would support batch processing and also the most accurate. Now, as we do not control all variables in the imaging process (users typically work with their own lighting studios) this, nor any method for background removal we explored will work flawless for every object. Output in JPG format allows users to choose any RGB color to use for the background (for ex. pure white). Output in TIFF or PNG format will retain transparent alpha background.
Below discusses the workflow, functionality, limitations and suggestions.
Shutter Stream 2.1 Update – We have introduce Chroma Key (green screen) Background Removal algorithms. Video tutorial is as follows. This will be a great addition for those using Shutter Stream for apparel photography on mannequins and models.
Automated Background Removal Workflow:
- Enable Live View mode then place an object in the field of view of the camera.
- Adjust camera settings to optimize for lighting environment (if required).
- Enable Crop and define Pre-Crop markers, that is the area that users wish to capture, around the subject (this can be defined at a perfect square or custom ratio if required).
- Capture image by clicking Snap button in software – immediately after image is captured, this will be uploaded into the Shutter Stream application for viewing.
- Repeat steps 1 – 4 while maintaining all variables consistent (to include Crop, Camera Settings, and Lighting).
- After product images are captured, user will be required to shoot a reference image – that is an image of just the background. Again, maintaining all previously used variables, user will want to toggle their camera from AF (auto-focus) to MF (manual focus) mode inside the software. This will ensure the camera focus does not fail when taking a background image.
- Select and import the Background Reference Image into the BR area in the editing mode of Shutter Stream
- After the Background Reference Image has been set, import the product images into the editing area, select the BR tool and adjust Threshold, Edge Sensitivity and Edge Blending to optimal values. As changes are made, they will be reflected in real-time in the Background Removal Preview Window. Users then have the ability to Apply the Background Removal changes to a single (Apply button) or entire set of images (Apply to All button).
- If required, users can take advantage of the other editing tools (Manual Cut Out Tool, Levels, Curves, Brightness, Saturation, Contrast, Sharpness & Hue) to make additional edits and color corrections.
- Output images using the Batch Saving or Dynamic Saving Tool (both tools include Batch Renaming, Resizing, Resolution Adjustments, Watermarking, Auto Edits and FTP Upload).
Automated Background Removal Tool Limitations:
As shadows are not present in the background only ‘reference image’ the Automated Background Removal Tool can have difficulties removing these from the frame. The harsher the shadow (darker) the more the user is required to increase Tolerance to remove the shadows. This can typically cause parts of the product to be removed – especially in cases in which the product does not contrast well from the background. Please see examples below:
Now, lets take a look at the results. For the image with the background we had to increase the tolerance in an attempt to remove the shadows underneath the product. We can see this starts to eat into the product slightly and there is still a very harsh shadow. With the image with the slight drop reflection (shoton an Iconasys LED LumiPad), we only had to make a minimal adjustment to remove the entire background and replace this with white:
- Use the Automated Background Removal Tool to cut out the majority of the background then use the Manual Cut Out Tool to manually cut out the rest of the shadow that was not removed. This solution will still be much more efficient than any 3rd party background removal tools.
- Use a bottom lit Iconasys LED LumiPad to shoot your products on. These inexpensive bottom lit LED Tables will eliminate harsh shadows, are available in many sizes to work in any lighting set up and were designed specifically for use with the Automated Background Removal Tool.
- Use a bottom and backlit Iconasys LED Light Box. These will ensure high quality results regardless of experience and were designed specifically for use with the Automated Background Removal Tool.
Users can also choose to manually cut out the shadow using the included Lasso Tool. So, the automated background removal tool will remove ~80% of the background, then users will simply clip the rest. This workflow is still much more efficient than removing backgrounds in other image editing applications. How it works video:
Products that don’t contrast well with their background:
The Automated Background Removal Tool can have a difficult time with products that don’t contrast well with their background color. As Tolerance is increased, it will remove pixels that differ more from the reference image. This will often lead to pixels in the product being removed. See example below:
Now, lets take a look at the results. For the product image on the left, shot inside the Iconasys LED Studio, we are able to create contrast between the object and the background color (Iconasys LED Studios include bottom and back lighting). So to remove the background, we only need a slight adjustment to the Tolerance. For the image on the right, we can see the object and background image have little contrast. This can cause a problem when increasing tolerance to remove the product background as it will start to remove parts of the product. Results are as follows:
- Shoot products on a contrasting background. If shooting white products (as we do suggest to shoot on a consistent white seamless backdrop) users will want to ensure the background is lighter than the product. This can be achieved by using an Iconasys LumiPad as the background (LED lit white panel to sit behind the object) then using less light on the object to achieve an underexposed image (as per example above). The same effect can be achieved using an Iconasys LED Lighting Studio. If shooting an underexposed image, Shutter Stream includes image editing tools that can be used for correcting colors after the background has been removed.
- Focus high wattage lights specifically on the white seamless background to blow it out. Then use less wattage lights on the product to ensure there is ample contrast between the product colors and the background colors.
Both options will also work well when using the Automated Background Removal Tool for glass objects.
Shutter Stream also includes the ability for users to add parts of the frame back in that might have been removed during the automated background removal process. Our ‘Alpha Opaque’ option allows users to use the Lasso tool to define a portion of an image to add back in. How it Works Video:
While the Automated Background Removal Tool in the Shutter Stream and Shutter Stream 360 Product Photography Software will provide a streamlined workflow for product photography background removal, it does have its limitations. It is important to shoot product images under correct lighting conditions to eliminate shadows and provide contrast between objects and the background. Should users not have the correct lighting conditions, the Automated Background Removal Tool will still prove to be a powerful tool to automate (in most cases) the majority of the background removal process. Remaining backgrounds can then be removed using the Lasso (manual cut out) Tool to achieve a complete white background. Shutter Stream and Shutter Stream 360 Product Photography Software workflows tools (camera control, image capture, image editing, image processing) is a sure way to save time, money and effort when shooting product images while enabling users, regardless of experience to create high quality results.