Tag: Nano Banana Pro

  • Nano Banana Pro Experiment: Spotting AI with the Naked Eye

    Nano Banana Pro Experiment: Spotting AI with the Naked Eye

    Key Takeaways

    1. High-Quality Image Generation: Nano Banana Pro produces nearly perfect images with accurate anatomy, consistent perspectives, and correct proportions, making it an impressive AI tool.

    2. Detail Inaccuracies: Despite its strengths, Nano Banana Pro can overlook small details, leading to noticeable mistakes when images are closely examined.

    3. Zipper and Fabric Issues: Specific problems include missing zipper teeth, inconsistent snap fasteners, and illogical collar designs, which reveal the AI’s limitations.

    4. Perspective Flaws: While the initial perspective appears correct, closer inspection shows inconsistencies in vanishing points and alignment, indicating potential flaws in the AI’s spatial understanding.

    5. AI Recognition Challenges: After minor edits, AI detection tools struggle to identify images as AI-generated, highlighting the importance of human judgment in recognizing subtle errors commonly found in AI-created images.

    Nano Banana Pro is a fantastic tool for creating stunning images with ease. Initially, the images produced by Google’s AI seem nearly perfect. The hands have four fingers and a thumb, and the anatomy and proportions look right. The perspectives are consistent, and the proportions are typically accurate. This image generator is so advanced that, after only minor post-processing, many automatic AI detection tools struggle, as shown in our first trial.

    Details Matter

    However, even Nano Banana Pro can miss some details. A little zooming in can expose various mistakes in the image. Let’s examine these issues closely.

    In our image, the woman is dressed in an olive-green jacket that has a style blending a field jacket and a parka. Nano Banana Pro captures the folds and fabric beautifully, but other aspects of the jacket are problematic. When we look closer, the inconsistencies become obvious.

    Zipper Problems

    Let’s focus on the zipper. On the right side of the image, the zipper teeth vanish halfway. The left side shows varying lengths and spacing of the zipper teeth. The snap fasteners reveal that this image was made by AI. Instead of round holes, we see a D-shape. Lastly, the jacket’s collar raises questions. The right side shows a white fur trim, which is missing on the left. Instead, the collar fabric merges smoothly with the shoulder pieces on the left. Moreover, the cut of the collar section displayed doesn’t make any sense.

    There’s also an issue with the zipper on the black fleece jacket. It lacks teeth, the pull tab is bent, and below, the zipper transforms into fabric.

    Perspective Flaws

    The devil is in the details here as well. Gemini, or Nano Banana Pro, does well with perspective at first glance. Only one vanishing point seems visible. But if you look closer, you can see that, despite the blur, some vanishing point lines cross different floors or, in the case of the two buses on the left, simply vanish into nothing. We highlighted the inconsistent lines in purple after identifying the perspective’s vanishing point.

    When it comes to hands, even Google’s top AI has slight flaws. If you examine closely, the fingers and hand of the woman holding the banana are not anatomically correct. There are clear issues with the middle finger. The width of the finger and fingernail is off, the front joint has no articulation, and the connection to the hand is unclear. Additionally, the skin folds between the fingers stretch down to the knuckles.

    At the hairline, it’s particularly noticeable that almost all the hair, including gray strands, starts with a brown base. This is especially striking at the front of her head.

    AI Recognition Challenges

    In the first part of our series, we put the image of the woman with the banana through several AI image recognition tools. After some editing, unfortunately, six online platforms declared that the image was not AI-generated. Relying on a human eye and common sense is much more useful in this case. The image shows a series of indicators that it is AI-generated. Anyone who remains doubtful should consider the inherent issues in generative image creation. Small details often aren’t rendered accurately, including the textures of fabrics, jewelry, knitwear, zippers, and other common items. Tire tread patterns and rims on cars and planes, wooden surfaces, and skin wrinkles frequently signal AI processing.

    Source:
    Link

     

    Tags: , ,
  • Nano Banana Pro Experiment: AI Detection Tools Fail to Deliver

    Nano Banana Pro Experiment: AI Detection Tools Fail to Deliver

    Key Takeaways

    1. Unregulated AI content in the legal field poses significant risks, highlighting the need for reliable detection tools.
    2. Current AI detectors struggle with new image generation models, creating blind spots that allow AI-generated images to evade detection.
    3. Basic modifications, like removing watermarks, can fool many detection tools, demonstrating their limitations.
    4. Advanced photo editing techniques can significantly reduce the detection probability of AI-generated images, making them appear as real photographs.
    5. The findings indicate that existing AI detection technologies are in their early stages and may provide a false sense of security in critical settings like courts and newsrooms.

    The case involving Melissa Sims has shown the significant risks posed by unregulated AI content within the legal field. There’s a promise in the form of “AI detectors,” which are software programs designed to cleverly determine if an image is created by a human or a machine. Still, doubts linger about how reliable these digital tools are, especially when they encounter intentional misleading. A test involving six of the most popular detection services aimed to evaluate their effectiveness.

    Experiment Setup

    To assess the performance of these detection tools, we employed the image generator known as “Nano Banana Pro,” based on Google Gemini. The prompt we chose was intentionally straightforward yet a bit playful: “A woman striking a fighting pose holding a banana against the sky. Background typical cityscape with uninterested people.”

    Challenges Faced by Detectors

    The first significant challenge for the detection software arose from the particular model in use. As Nano Banana Pro is relatively new, many detectors are not equipped to recognize it, presenting a blind spot.

    These detection services often depend on machine learning techniques and are trained to spot the specific “fingerprints” of established models like Midjourney, DALL-E 3, Stable Diffusion, or Flux. Thus, a novel model like Nano Banana Pro has a distinct edge since its specific generation patterns are not yet included in the detectors’ training datasets, making it easier to evade detection.

    Initial Testing Results

    Initially, the detectors had a straightforward job. The original PNG file was converted into a JPG, and its metadata got wiped. Notably, though, the image still displayed a clear Gemini watermark.

    One might think that a visible AI watermark would be easy for detection software to catch. However, the outcome was surprising: even in this basic form, two out of the six tools failed to identify the image as AI-generated, despite the clear watermark.

    Realistic Forgery Scenario

    Next, we moved toward a more realistic forgery scenario by removing the identifying watermark. Using the AI eraser in the default Windows Photos app, this change was made seamlessly in just a few seconds.

    After this simple edit, the results shifted. One more tool was fooled, now indicating a low probability of AI generation. Interestingly, Illuminarty actually raised its probability rating for the image being AI-generated post-edit. Nevertheless, three of the six tools still estimated less than 30% likelihood that the banana-holding woman was an AI creation.

    Final Testing Phase

    The final test was crucial. AI-generated images often appear “too smooth” and perfect, lacking typical noise. To effectively mislead the detection tools, the image needed to incorporate artificial “realism,” simulating common errors found in digital photography. Using Cyberlink PhotoDirector, adjustments were made. Lens correction was applied, artificial chromatic aberration created color fringes, contrast was boosted, and realistic image noise was added. The goal was to make the image resemble a photo taken by an imperfect camera—all accomplished in just a few minutes.

    The outcome of this last phase was a complete triumph over the detection technology. After undergoing standard post-processing, all six detection services failed to assert that the image was AI-generated, with none indicating a probability greater than 5 percent. For the software, the image of the woman with the banana was now unmistakably a real photograph.

    Conclusion

    Our experiment clearly illustrates that current AI detection technologies are still very much in their early stages. If it only takes a few minutes with standard photo editing tools to reduce detection rates from “very likely” to “under 5 percent,” these tools are not just ineffective for courts, newsrooms, or law enforcement—they pose a risk. They foster a false sense of security that simply is not warranted. The idea of “trust, but verify” only holds if the verifiers aren’t blind.

    Source:
    Link

     

    Tags: , ,