Shop here: https://tf.style/collections/all?filter.v.price.gte=&filter.v.price.lte=&filter.p.tag=Anti&sort_by=price-descending

Anti-Surveillance At the heart of this new collection is a simple question: how can we translate cutting-edge adversarial research into eye-catching,-style-forward garments? We surveyed the academic literature on “adversarial fashion” and honed in on two core insights:

Macro-Scale Disruption Most “anti-surveillance” clothing on the market relies on tiny, high-frequency noise or pixel-level glitches. Those patterns vanish under real-world blur—at 10 m or more, a camera simply sees a gray blob.

In contrast, our designs employ large shapes and blocks of high-contrast color (neon, pastels, stark black/white). At typical CCTV distances, these large edges still register as conflicting contours, forcing modern detectors to lose confidence in their bounding-box output.

High-Contrast, Irregular Placement Rather than neat, grid-aligned prints, we intentionally offset each motif so detectors cannot anchor to repeatable symmetry. That irregularity further fragments your silhouette.

Our color palettes pair deep dark fields with vibrant accents of cyan, magenta, neon yellow, or pastel gradients. Under auto-exposure, those bright hues “bloom” on many cameras, amplifying edge noise.

What We Didn’t Do (and Why) Small-Scale Speckle Noise: We avoided 1–2 cm “static” noise prints. In lab demos, such fine detail only fools a detector at 2–3 m. Beyond that, the pattern becomes indistinguishable from ordinary texture.

Face-Focused Dazzle or Makeup: Our goal was to disrupt the whole-body detector, not just a facial-recognition API. We did not incorporate paint-on masks or IR-LED glasses, which can only hide a face under very specific angles and lighting.

Thermal or IR-Reflective Materials: Some researchers embed aerogel or metallic threads to confuse infrared sensors at night. Those approaches require specialized fabrics and only work in low-light/NIR mode; we opted for standard sublimation-print processes on poly blends.

By concentrating on large, high-contrast shapes that tile seamlessly across a garment, our line diverges from many “anti-surveillance” brands that rely solely on style-transfer textures or catwalk-style “CV Dazzle.” The result is a closet of bold, streetwear-ready pieces that also happen to harness adversarial theory.

Important Caveats & Disclaimers Untested in the Wild. To date, none of these patterns have undergone formal field trials, or testing of any kind, against production-grade multi-camera networks, thermal fusion systems, or adversarially robust models (e.g., YOLOv8, DETR with adversarial training).

Obligations & Local Regulations In some jurisdictions, deliberately obscuring one’s image or attempting to evade lawful surveillance may be restricted or penalized, especially in sensitive areas (airports, banks, public transit). Always check local regulations before purchasing or wearing these garments.

No Warranty of “Invisibility” Wearing one of our patterns may reduce detection confidence by a few percentage points on certain camera-model/lighting combinations, but it does not guarantee you’ll “vanish” from view. Human eyes, multi-view cameras, and retrained models can still identify wearers.

Our mission is to blend disruptive technology with everyday style. We’ve taken the core adversarial insights—large-scale edge disruption and irregular, high-contrast layouts—and wrapped them in fashion-forward prints that stand out in a crowd.