Cropping Labels

Introducing Cropping Labels in Release 25.08.07

With the release of 25.08.07, we’re excited to introduce a new set of Cropping Labels designed to help RemotePhoto Butler deliver more consistent, specific feedback based on your organization's photo cropping standards.

These labels function just like all other labels in the HelperBot Explainability system and are fully compatible with Smart Filters. Once enabled, they allow your Butler to better identify whether a submitted photo aligns with your institution's framing preferences.

Note: These new labels do not alter or control how a photo is cropped. Instead, they are used purely for feedback purposes—helping your users understand if their photo meets your established standards for proper positioning and framing.

New Cropping Labels: What They Are

Each cropping label evaluates whether the subject in the photo (usually the cardholder’s face) is framed correctly according to specific alignment criteria. These labels operate based on a series of classifiers, each producing a confidence score indicating how likely it is that a certain condition is present.

Here are the cropping labels introduced in this release:

Label	Description
Face	The "Face" label (like its sibling labels, "Head" and "Person") is part of the cropping classifier’s output and helps determine whether a user's face is properly framed within a photo. Each variation focuses on a specific dimension or boundary of the detected face. Each attribute is calculated using a normalized coordinate system ranging from 0 to 100, with `0` representing the top edge of the image and `100` representing the bottom. Height - Represents the vertical span of the face within the frame, from top to bottom. Width - Measures how much horizontal space the face occupies in the photo. Top - Indicates the vertical position of the top of the face (typically aligned with the hairline). Calculated on a 0–100 scale from top to bottom of the image. Bottom - Indicates the vertical position of the bottom of the face (typically the chin). Also calculated on the 0–100 top-to-bottom scale. Left - Measures how far the face extends to the left side of the image. Useful for evaluating horizontal centering. Right - Measures how far the face extends to the right side of the image. It complements the “Left” label for centering and width analysis.
Head	Height - Represents the vertical span of the head (typically recognized as the top of someone's hair) within the frame, from top to bottom. Width - Measures how much horizontal space the head occupies in the photo. Top - Indicates the vertical position of the top of the head (typically aligned with the top of a person's hair, not the hairline). Calculated on a 0–100 scale from top to bottom of the image. Bottom - Indicates the vertical position of the bottom of the head (typically the chin). Also calculated on the 0–100 top-to-bottom scale. Left - Measures how far the head extends to the left side of the image. Useful for evaluating horizontal centering. Right - Measures how far the head extends to the right side of the image. It complements the “Left” label for centering and width analysis.
Person	Height - Represents the vertical span of the Person Box. Width - Measures how much horizontal space the Person Box consumes in the photo. Top - Indicates the vertical position of the top of the Person Box. Calculated on a 0–100 scale from top to bottom of the image. Bottom - Indicates the vertical position of the bottom of the Person Box. Also calculated on the 0–100 top-to-bottom scale. Left - Measures how far the Person Box extends to the left side of the image. Useful for evaluating horizontal centering. Right - Measures how far the Person Box extends to the right side of the image. It complements the “Left” label for centering and width analysis.
Brow Height	This label identifies the vertical position of the brow line—a consistent anatomical reference point—within the photo frame. It can be used to assess how high or low the subject's head appears relative to the top edge of the image. The Brow Height label references this position on a normalized scale, supporting precise calculations of head alignment. This helps ensure the photo aligns with established cropping standards, such as those used in official ID specifications. When combined with other cropping labels, this metric provides stable and reliable feedback on head placement, making it easier to guide users toward properly framed submissions.
Eye Height	The Eye Height label captures the vertical position of the subject’s eyes within the frame, based on a normalized coordinate scale from 0 to 100 (top to bottom of the image). This metric serves as a stable reference point for evaluating the subject’s head alignment in relation to industry-standard photo composition guidelines. Because eye position is one of the most consistent features across faces—and a key anchor in official ID photo requirements—this label helps determine whether the subject's face is positioned too high, too low, or appropriately centered. When combined with other facial position metrics (like Brow Line, Chin, and Face Height), the Eye Height label contributes to a more accurate and consistent feedback loop, allowing RemotePhoto Butler to guide users toward optimal framing without guesswork or trial and error.
Chin Height	The Chin Height label identifies the vertical position of the subject’s chin within the photo, calculated using a normalized 0–100 coordinate scale (with `0` representing the top of the image and `100` the bottom). This measurement helps determine how low the head is positioned in the frame and plays a key role in verifying photo alignment. Because the chin represents the lower boundary of the face, this label is particularly useful when assessing whether the head is fully visible and proportionally framed. In combination with Eye Height, Brow Line, and Face Top, the Chin Height label allows RemotePhoto Butler to provide more detailed and accurate feedback about the vertical placement of the face. It also helps detect common issues like: Face being cropped too tightly at the bottom Excessive spacing between the chin and the bottom edge of the image Face appearing too low in the frame overall This label contributes to a more precise evaluation of photo compliance with official ID and institutional standards, improving both automation and user guidance.
Vertical Center Line	The Vertical Center Line label measures the horizontal alignment of the subject’s face within the frame. It determines how far left or right the vertical midpoint of the face appears, based on a 0–100 coordinate scale, where: `0` represents the far left edge of the image `100` represents the far right edge of the image This label is essential for evaluating whether the subject is centered properly—a common requirement in ID photo standards. Default interpretation Guidelines: A score below 45% suggests the subject is too far to the left A score above 55% suggests the subject is too far to the right Scores between 45–55% are generally considered well-centered By focusing on the vertical midpoint of the detected face or head, this label helps RemotePhoto Butler deliver consistent, quantitative feedback about centering—without relying on visual guesswork. It is particularly effective when paired with other labels like Face Width or Face Not Centered to reinforce confidence in centering assessments. This label can also trigger actions within Smart Filters, such as recommending a resubmission when a face is off-center beyond the acceptable range.

Each label is attached to the photo along with a confidence score ranging from 0.0 to 1.0, where a higher score indicates stronger evidence that the issue is present.

Want an example?

Example: Aligning Cropping Labels with U.S. Passport Photo Standards

One of the most practical applications of Cropping Labels and Smart Filters is the ability to align your institution’s photo requirements with established national or industry standards. A great example of this is configuring RemotePhoto to enforce the U.S. Department of State’s guidelines for passport photos.

While the Department of State describes measurements in relation to the bottom of the image, RemotePhoto uses a coordinate system measured from the top (with 0 at the top and 100 at the bottom). As a result, these values are inverted when configuring Smart Filters.

Here’s how you might set up label-based triggers to match those passport photo requirements:

Recommended Smart Filter Configuration

Auto Deny if Head Height is less than 50% or greater than 69%
This ensures the head fills an appropriate portion of the photo—neither too small nor too large.
Auto Deny if Eye Height is less than 31% or greater than 44%
The Department of State requires the eye line to fall between 1 1/8" and 1 3/8" from the bottom of the image. When converted to our coordinate system (from the top), this corresponds to a vertical range of 69% to 56%, or 31% to 44% from the top.
Auto Deny if Vertical Center Line is less than 45% or greater than 55%
This ensures the face is horizontally centered, allowing only a slight margin of left/right positioning.

Why This Matters

By configuring these Smart Filters with Cropping Labels, your photo intake process can automatically enforce strict standards with no manual review required—unless an image falls into a gray area. This helps reduce human workload, improve consistency, and ensure your submitted photos comply with widely accepted criteria.

Even if your organization doesn’t require U.S. passport-level precision, this example illustrates how flexible and powerful the label infrastructure is when paired with Smart Filters.

What Cardholders Will See

When a Cropping Label is triggered, the end user may receive more informative feedback such as:

“It looks like your face is too far away in the photo. Try standing a bit closer and make sure your face fills the center of the frame.”

This goes a step beyond traditional rejections that simply say “photo doesn’t meet requirements.” Now, users know why and how to get it right on the next try—reducing frustration and repeat submissions.

Release Video

Need Help Getting Started?

Our support team is happy to assist you in setting up Cropping Labels, customizing Smart Filters, or reviewing label performance across your organization.

Reach out to us anytime at support@cloudcard.us, and we’ll help make sure your RemotePhoto Butler is configured just the way you need.