Yalantis/uCrop · error · CImgArgumentException

displacement(): Invalid specified precision %g (should be…

Error message

displacement(): Invalid specified precision %g (should be >=0)

What it means

displacement() takes a precision parameter that controls the convergence threshold of the iterative optical-flow solver. Negative values are meaningless for this stopping criterion, so the library throws this CImgArgumentException when precision<0.

Solutions

  1. Pass a positive precision (the default is 5.0f); lower values give more precise but slower convergence
  2. Clamp the value before calling: if (precision < 0) precision = 0 (or a sensible default)
  3. Validate user/config input with a range check (precision >= 0) at parse time
  4. Check code that computes precision dynamically for sign errors

Example fix

// before
img.displacement(ref, 0.1f, precision /* -1.f sentinel */);
// after
img.displacement(ref, 0.1f, precision < 0 ? 5.0f : precision);
Defensive patterns

Strategy: validation

Validate before calling

float precision = cfg.precision;
if (precision < 0) precision = 5.0f; // library default
CImg<float> flow = img.displacement(reference, smoothness, precision);

Try / catch

try {
  CImg<float> flow = img.displacement(reference, smoothness, precision);
} catch (CImgArgumentException& e) {
  std::fprintf(stderr, "precision must be >=0: %s\n", e.what());
  CImg<float> flow = img.displacement(reference); // default precision=5
}

Prevention

When it happens

Trigger: Calling displacement() with a negative precision argument, e.g. displacement(ref, smoothness, -1.0f), typically from a miscalculation, a sign error, or an unvalidated user/config value.

Common situations: Config files where precision was meant as an offset or tolerance and a negative value was entered, variables initialized with -1 as a sentinel and passed unmodified, or unit confusion (pixels vs subpixels).

Related errors

AI-assisted analysis of Yalantis/uCrop@f788b534b4 (2026-09-08). Data as JSON: /api/errors/a334235c57d362fc. Report an issue: GitHub.

Appendix: source

Thrown at ucrop/src/main/jni/CImg.h:45966

        move_to(*this);
    }

    //! Estimate displacement field between two images \newinstance.
    CImg<floatT> get_displacement(const CImg<T>& reference,
                                  const float smoothness=0.1f, const float precision=5.f,
                                  const unsigned int nb_scales=0, const unsigned int iteration_max=1000,
                                  const bool is_forward=false,
                                  const CImg<floatT>& guide=CImg<floatT>::const_empty()) const {
      if (is_empty() || !reference) return +*this;
      if (!is_sameXYZC(reference))
        throw CImgArgumentException(_cimg_instance
                                    "displacement(): Instance and reference image (%u,%u,%u,%u,%p) have "
                                    "different dimensions.",
                                    cimg_instance,
                                    reference._width,reference._height,reference._depth,reference._spectrum,
                                    reference._data);
      if (precision<0)
        throw CImgArgumentException(_cimg_instance
                                    "displacement(): Invalid specified precision %g "
                                    "(should be >=0)",
                                    cimg_instance,
                                    precision);

      const bool is_3d = reference._depth>1;
      const unsigned int spectrum_U = is_3d?3:2;

      if (guide &&
          (guide._width!=_width || guide._height!=_height || guide._depth!=_depth || guide._spectrum<spectrum_U))
        throw CImgArgumentException(_cimg_instance
                                    "displacement(): Specified guide (%u,%u,%u,%u,%p) "
                                    "has invalid dimensions.",
                                    cimg_instance,
                                    guide._width,guide._height,guide._depth,guide._spectrum,guide._data);
      const float
        scale_factor = 2.f,
        abs_smoothness = cimg::abs(smoothness),

View on GitHub (pinned to f788b534b4)