EyeQ

What are EyeQ and Perfectly Clear?

EyeQ Imaging Inc. is a Canadian company specializing in imaging technology; its core technology is Perfectly Clear, an AI-based system for automatic correction of images and videos. Its products are designed for use in photo printing, school and sports photography, tourism-related applications, imaging software, as well as platforms for handling large volumes of content.

It is not a general-purpose programming assistant; rather, it is an image processing engine that can be integrated into business systems, clients, and camera workflows. It also offers cloud-based Web APIs that require no local installation.

Product format

Main functions

Intelligent automatic color calibration

Perfectly Clear analyzes the scene, exposure, faces, skin tone, and color conditions, then selects the appropriate correction parameters. Users can opt for AI Preset Selection or use fixed presets to ensure consistency in the style of a batch of photos.

• Exposure and face detection exposure: Adjusts the overall brightness while preventing the faces from being affected by the background.
• White balance and color shift removal: Addresses unnatural cool or warm colors as well as abnormal color deviations.
• Color restoration and vibrancy: Enhance the colors while striving to maintain a natural appearance.
• Contrast, Depth, and Sharpening: Improve levels, details, and visual clarity.
• Noise reduction and red eye correction: Reduces noise at high ISO levels and fixes noticeable red eyes.

AI applications and skin tone recognition

Scene detection is used to assign photos to the most appropriate predefined categories, while AI Skin Tone identifies different skin tones and makes adjustments to correct them. This feature is suitable for tasks that require consistency, such as school portraits, event photography, and group photos.

Portrait retouching

Retouching 2.0 and its related parameters allow for adjustments to skin, eyes, teeth, and various facial details. Developers can control the intensity of these adjustments according to the requirements of a particular brand, rather than being limited to pre-set filters.

• Smooth skin, removal of imperfections, and control of facial oiliness.
• Enhanced eyes, enlarged eyes, and reduced dark circles under the eyes.
• Teeth whitening, as well as portrait depth and volume.
• Some old parameters are marked as no longer in use in the current SDK; when integrating, it is necessary to migrate to the latest version of those fields.

Automatic cropping, scaling, and format conversion

Web APIs can automatically crop and adjust the output dimensions based on preset settings, and they can also convert the output format. SDKs and command lines are suitable for integrating cropping coordinates or fixed presets into existing production processes.

AI amplification

AI Upscaling adds pixels to low-resolution images and increases their output size. It is a form of generative capability useful for enlarging small images, but it cannot guarantee the restoration of real details that were not present in the original image.

Background removal and replacement

Background removal is achieved by using segmentation models to distinguish between the foreground and the background; cloud-based services also allow for the replacement of the background with images, solid colors, or AI-generated backgrounds. However, the edges of people, transparent objects, and complex hair patterns still require manual inspection.

Anti-reflective coating for glasses

The system can detect and correct reflections from glasses that obscure the eyes, with generative AI being used for the content restoration process. The results of this correction may alter the details of the eyes; it should not be used for purposes such as identification, medical use, or evidence collection without further verification.

Video correction

Web APIs and cloud-based services support video enhancement, with pricing based on resolution and the number of processing seconds. Video processing is useful for standardizing exposure, color, and overall appearance, but the time required for uploading and the processing costs increase as the file size, resolution, and duration increase.

Web API workflow

1. Create a Perfectly Clear cloud account to obtain an API key, and first use test materials that do not contain sensitive information.
2. Upload JPEG, PNG, MP4, or MOV files, and record the file identifier returned by the service.
3. Submit the correction task, select an AI preset or a custom preset, and set the cropping, scaling, background, or reflection parameters as needed.
4. Save the task identifier, check its status via polling, or set up a Webhook to receive completion notifications.
5. After the task is completed, download the correction results; do not use the temporary public address as a long-term storage location.
6. Check the portrait, edges, skin tone, exposure, and output size, then send the results to the subsequent delivery process.

Process for integrating with the local SDK

1. Confirm with EyeQ the Server, Desktop, Mobile, or Browser SDKs, as well as the target platforms and the number of authorization nodes.
2. Obtain the SDK package, license, certificates, AI model files, and the corresponding version documentation.
3. Create and reuse processing engines, and load AI scenarios as well as color correction, skin tone, and face recognition models.
4. Pass the RGB image buffer or file object to the analysis stage to obtain profiles of the scene, faces, and so on.
5. Load the preset exported from Workbench, or set the intensity for exposure, portrait mode, noise reduction, and so on within the parameter object.
6. Apply corrections, save the results, and release the profile, image, and engine resources.
7. Use representative data to test speed, memory, threads, authorized nodes, and abnormal image processing.

Input/Output and File Limits

Workbench also supports RGB 8-bit or 16-bit JPG, TIFF, BMP, and PNG files, and it can load files in their raw format using its RAW engine; HEIC files are not supported at the moment. The format support varies depending on the specific product.

Performance and deployment options

• The cloud-based Web API requires no installation and enables the simultaneous sending of numerous tasks, but it involves uploading and downloading original files.
• The Server SDK and Desktop SDK run locally, with latency and throughput depending on the client’s hardware.
• Docker is suitable for placing similar APIs within a private network, so that the images remain within the customer’s environment.
• The Browser SDK performs the main processing in the browser, making it suitable for instant preview; however, it still requires license verification and domain binding.
• The SDK recommends that each thread use one engine and reuse it across multiple images, in order to avoid the overhead associated with repeated initialization.
• The SDK node license counts virtual machines and cloned instances within the total number of active instances; it is necessary to verify the authorization model before scaling out.

Which users are it suitable for

• Schools, sports, and event photography agencies: unified exposure, skin tone adjustment, portrait processing, and cropping on a bulk basis.
• Photo printing and ordering platform: Automatically improves the quality of images throughout the process of uploading, previewing, and placing orders.
• Tourist attractions and cruise ship photography: High-volume processing of numerous photos taken by tourists.
• Image software developers: integrate correction engines into desktop, mobile, or web-based products.
• E-commerce and content platforms: They handle products, portraits, and images uploaded by users, but manual inspection is required.
• Organizations with strict privacy requirements: Choose SDK or Docker to keep the images within the local network.

Price packages

The Web API tracks usage on a monthly basis, with each group of 100 images serving as a billing unit. The table below shows the current public prices; if the usage exceeds these limits or local deployment is required, you should contact sales.

Photo resize price

Perfectly Clear Cloud features a separate resizing option and a pricing structure based on increments of $10; its costs cannot be combined with those of the Web API. Before placing an order, it is necessary to confirm the currency used, any taxes or fees, the minimum monthly fee, and any corporate discounts.

Trial, cancellation, and refunds

• The Web API and Cloud currently offer 100 photo corrections and 5 minutes of video testing.
• QuickServer offers a 15-day fully functional trial without watermarks, and an enterprise email address is required.
• The Workbench product page offers a 30-day trial period for previewing corrections and creating presets.
• If a Web API is canceled mid-month, the change takes effect immediately; no further charges will be applied, but any unused volume still needs to be paid for.
• It can be downgraded to the Free tier; the terms do not provide a general guarantee of proportional refund.
• Requests to delete an account will be processed; according to the terms, this will not happen earlier than 10 business days after the request is received.

Product advantages

• The same correction technology can be deployed in the cloud, on local servers, on desktops, on mobile devices, and in browsers.
• Automatic scenes and presets are suitable for bulk, uniform processing, while still allowing for parameter customization.
• It covers not only color correction but also portraits, cropping, zooming, backgrounds, reflections, and videos.
• The local SDK and Docker offer the option of not transmitting images outside the network for privacy-sensitive workflows.
• The document contains examples for C, .NET, Java/JNI, Android, Browser SDK, and Web API.

Capacity boundaries

• Automatic correction is affected by the exposure, compression, resolution, color space of the original image, as well as the size of the face.
• Thumbnails may reduce the quality of scene, face, and skin tone detection; it is advisable to use input at full resolution before reducing the size.
• Generative amplification and reflection restoration can create details that are not present in the original image, making them unsuitable for tasks that require evidence fidelity.
• Background segmentation may suffer from edge errors in cases of hair, transparent objects, motion blur, and backgrounds with similar colors.
• The SDK has requirements regarding node activation and processor instructions; version 11 requires AVX2, and it is necessary to verify this before deploying it on older hardware.
• The marketing page states that the Web API offers a 99.5% availability guarantee, but the official terms specify that no guarantee is provided for continuous availability or reliability; contract clients should adhere to the service level agreements that have been signed.

Privacy and data storage

• The original images and results from the Web API are accessible via temporary public addresses; they remain available for at least 24 hours and no more than 7 days, after which they are automatically deleted.
• The terms state that Customer Content is cached for 7 days without any off-site backup, so customers should download it promptly and create their own backups.
• The data may be processed in the United States and Canada, and customers are responsible for complying with cross-border and applicable privacy regulations.
• Do not upload identification documents, bank account numbers, or any sensitive information subject to special data security laws.
• EyeQ employees generally do not access images; they can view them only when customers explicitly request support or diagnosis.
• The website collects data on sales, purchases, support, and subscriptions, and it uses cookies; corporate licensing statistics may include license keys as well as aggregated usage figures.
• The privacy policy states that data is not sold or rented, and that only the information necessary is shared with partners who provide essential services such as order processing, email support, and assistance.

AI training and exit options

The Web API terms permit the use of Customer Content to develop, improve, and train products, provided that the user agrees to data sharing; an option to withdraw consent is available in the account settings. Withdrawing consent affects only future uploads and does not affect data that has already been shared.

EyeQ’s AI policy specifies that customer images will not be used for training without consent, and images deployed locally are not sent back to EyeQ. Enterprise customers should clarify in the contract the default settings, the time at which usage can be stopped, and the products applicable to this arrangement.

Content ownership and commercial use

The customer retains ownership of the images, videos, and correction results uploaded. To provide the service, the customer grants EyeQ permission to host, store, copy, modify them, and carry out any necessary derivative processing.

• The customer must possess the rights to the content and obtain the necessary consent from the subjects, employees, or end users.
• The Perfectly Clear software, SDKs, documentation, and algorithms remain the property of EyeQ; purchasing a license does not grant access to the source code.
• Generative outputs such as background replacement, AI Art, zooming, and reflection correction still require checks regarding copyright, privacy of individuals, and authenticity.
• The scope of commercial use is determined by the SDK, API, or enterprise contract signed; a free trial cannot replace a production license.

GitHub, APIs, and open-source status

Pre-launch checklist

1. Choose between a cloud-based or on-premises solution, and determine the desired throughput, latency, privacy requirements, and budget.
2. A test set is created using the most challenging photos from real-world scenarios, covering various skin tones, glasses, backlighting, noise, and complex backgrounds.
3. Confirm the tiered pricing, resolution ranges, minimum billing units, and corporate discounts for each feature.
4. Define the default values for data sharing, training termination, 7-day caching, cross-border processing, and access procedures.
5. Verify that the output ICC, EXIF, orientation, thumbnail, format, and size meet the requirements of the downstream system.
6. Set up sampling checks, fallback procedures in case of failures, and manual correction processes for automatic correction.
7. When deploying locally, check AVX2, memory, threads, container resources, and the number of license nodes.

Frequently Asked Questions

What is the relationship between EyeQ and Perfectly Clear?

EyeQ Imaging Inc. is the development and operation company, while Perfectly Clear is its brand for technologies and products used for the automatic correction of images and videos.

Is Perfectly Clear available only through the cloud API?

No, it also provides servers, desktop, mobile, and browser SDKs, as well as Docker, the command line, QuickServer, and Workbench.

Can the Web API be tested for free?

Yes, the current testing quota allows for 100 photo corrections and 5 minutes of video processing, and no bank card is required; actual usage will be charged based on the specific functions employed.

How long are uploaded images stored?

The original images and results from the Web API are accessible for at least 24 hours and up to 7 days, after which they are automatically deleted; the platform does not provide any offline backups.

Will the customer’s images be used to train AI?

The terms provide options for data sharing and withdrawal; the AI policy states that data will not be used for training without consent. Users should confirm the default settings in their account and contract, and withdrawal does not affect any data that has already been shared.

Is Perfectly Clear open source?

No, the core engine, SDKs, and cloud services are subject to commercial licensing; what is available under the MIT license are only some examples of Web API calls, not the processing engine itself.

Can AI image editing completely replace human work?

It’s not suitable; while batch automation can improve efficiency, tasks such as handling portraits, the edges of backgrounds, generative zooming, and reflection correction still require random inspections and manual adjustments where necessary.

What is the price of the SDK?

SDKs, Docker, and most on-premises products are available through sales inquiries; the price depends on the platform, nodes, licenses, usage volume, and terms of the contract.

Summary

Perfectly Clear is suitable for imaging applications that require consistent batch color correction, portrait enhancement, and integration across multiple platforms. The Web API enables quick integration, while the local SDK and Docker offer greater control and better privacy protections.

When making a selection, it is necessary to take into account factors such as cumulative billing, file size limits, 7-day caching, termination of training processes, authorized nodes, and automatic error correction, rather than focusing only on the basic price per item.